{"id":"469e14f8-2059-4407-bb2d-04f1d3e978f1","arxiv_id":"2507.04467","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The Bilinear Hilbert-Carleson operator along the moment curve (t, t^2, t^3) satisfies the expected L^{p1} x L^{p2} to L^r bounds for 1 < p1, p2 < infinity and 1/2 < r < infinity.","lead":"This paper proves that a curved version of the bilinear Hilbert-Carleson operator, with inputs along t and t^2 and phase t^3, is bounded on L^p spaces for the full expected exponent range. The result settles the purely non-zero curvature case of a program connecting Carleson's theorem with the trilinear Hilbert transform.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Main Theorem 1.1 is only proved for the model parameters a=(1,-1,1), alpha=(1,2,3); the asserted reduction to this case is not demonstrated and is not a simple reparametrization for arbitrary pairwise distinct alpha.","rationale":"The paper contains a substantial and plausibly correct proof for the model case a=(1,-1,1), alpha=(1,2,3), but the advertised Main Theorem 1.1 covers arbitrary pairwise distinct nonzero alpha_j and nonzero a_j. The reader's weakest assumption identifies exactly this gap: the reduction from the general case to the model case is asserted, not proved. My stress-test confirms that this is the most load-bearing concern. A direct algebraic check shows that no simple change of variables can reduce generic exponent triples to (1,2,3), and the proof's dyadic phase analysis relies on the specific ordering and positivity of the exponents (e.g., |t|^3 <= t^2 <= |t| for k >= 0). Therefore the full theorem is not established by the present manuscript. However, the special-case result may well be valid and the gap may be addressable by either proving the claimed versatility or restricting the statement; hence I recommend keeping the reader's CONDITIONAL verdict rather than escalating to REJECT. The concrete test proposed would settle the issue by checking whether the asserted reduction exists and whether the key phase estimates survive for non-model exponents.","tokens_in":77301,"tokens_out":3761,"duration_ms":46145,"concrete_test":"Determine whether there is any bijective change of variables t = phi(s), preserving the principal value and the supremum in lambda, that transforms BHC_{a,alpha} into the model operator with alpha=(1,2,3). This requires solving a_j phi(s)^{alpha_j} = c_j s^j for j=1,2,3; show this forces alpha_j proportional to j. Then, for a concrete non-proportional triple such as alpha=(-1,-2,-3), trace the proof of Sections 2.2 and 3.1 to check whether the phase derivative analysis and the 'low oscillatory' versus 'stationary' classification still hold. If either the reparametrization fails or the classification breaks, Main Theorem 1.1 is unproven in full generality.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim (Main Theorem 1.1) asserts boundedness for all nonzero a_j and all pairwise distinct nonzero alpha_j. The proof, however, is carried out only for a=(1,-1,1), alpha=(1,2,3): Section 2 states 'for the reminder of the paper we set vec a=(1,-1,1), vec alpha=(1,2,3) and focus only on the corresponding BHC.' The only justification for the full statement is the assertion in Section 1.3 that 'the versatility of our methods allows us to treat with minimal effort any nonresonant choice of vec alpha' and that 'it is enough to only focus on a special case.' This transfer is never supplied. It is also not a trivial reparametrization: a change of variables t = phi(s) mapping a_j phi(s)^{alpha_j} to c_j s^j for j=1,2,3 would require alpha_1/1 = alpha_2/2 = alpha_3/3, so generic triples such as alpha=(-1,-2,-3) or (1,2,-1) are not equivalent to the model. Moreover, the proof uses in an essential way the ordering of powers, e.g. the inequalities |t|^3 <= t^2 <= |t| for |t| <= 1 (Section 3.1) and the stationary-phase classification in Sections 2.2 and 5; for negative or unordered exponents these relations fail and the same decomposition into BHC^Lo, BHC^Delta, BHC^notDelta is not justified. Thus the strongest claim as stated is not established, even though the special-case result may be correct.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims the maximal (up to endpoints) L^{p1} x L^{p2} -> L^r boundedness range for the purely non-resonant Bilinear Hilbert-Carleson operator BHC_{[a,alpha]} along monomial curves, for all nonzero a_j and all pairwise distinct nonzero alpha_j, in the Holder range 1/p1+1/p2=1/r with 1<p1,p2<infinity and 1/2<r<infinity. The proof is carried out only for the model case a=(1,-1,1), alpha=(1,2,3), via a decomposition into low oscillatory, non-stationary off-diagonal, stationary off-diagonal, and diagonal components, and relies on the Rank II LGC method and on the authors' prior work [24], especially Proposition 1.18 (= [24, Theorem 4.3]). The paper contains detailed arguments for the k>=0 regime of the diagonal and stationary components, with some central statements (Theorem 4.1, Theorem 6.1, parts of Proposition 3.4 and Section 3.7, and the entire k<0 appendix) only sketched or deferred.","tokens_in":77615,"tokens_out":3706,"duration_ms":43610,"significance":"If the proof is completed, the result is significant: it would settle the maximal boundedness range for this class of curved bilinear Carleson operators, complementing the companion treatment of the curved trilinear Hilbert transform in [24] and constituting a nontrivial application of the Rank II LGC method in a quasi-Banach range. The conceptual decomposition and the constancy-propagation strategy are valuable and are presented with considerable detail. However, the paper's strongest claim, Main Theorem 1.1 for arbitrary a and alpha, is not supported by the proof as written: the reduction to the model case is asserted but not demonstrated, and several load-bearing estimates are deferred. The significance of the full statement therefore depends on closing these gaps.","major_comments":[{"comment":"Main Theorem 1.1 is stated for all a_j in R\\{0} and all pairwise distinct alpha_j in R\\{0}, but Section 2 explicitly fixes a=(1,-1,1) and alpha=(1,2,3) for the remainder of the paper. The only transfer argument is the assertion in Section 1.3 that 'the versatility of our methods allows us to treat with minimal effort any nonresonant choice of alpha' and that 'it is enough to only focus on a special case.' This is not a proof, and it is not a trivial reparametrization: a change of variables t=phi(s) mapping a_j phi(s)^{alpha_j} to c_j s^j for j=1,2,3 would require alpha_1/1 = alpha_2/2 = alpha_3/3, which fails for triples such as alpha=(-1,-2,-3) or (1,2,-1). Moreover, the proof uses in essential ways the ordering of exponents, e.g. the inequalities |t|^3 <= t^2 <= |t| in Section 3.1 and the stationary-phase classification in Sections 2.2 and 5, which do not hold for negative or unordered exponents. Consequently, the paper establishes at most the special case a=(1,-1,1), alpha=(1,2,3); the full statement of Main Theorem 1.1 requires either a complete reduction argument or a restriction of the theorem's scope.","section":"Sections 1.3 and 2; Main Theorem 1.1"},{"comment":"Theorem 4.1 is a load-bearing estimate for the stationary off-diagonal component BHC^{notDelta,S}, specifically for Case (O2), and it is also used in Section 6 to obtain the full quasi-Banach range. Its proof, however, is a single paragraph asserting that it 'follows by a straightforward modification' of the proof of Theorem 3.1, with details left to the reader. The three listed ingredients are not enough to verify the claim: the constancy-propagation argument in the regime of Theorem 4.1 has additional frequency-localization suppositions and a renormalized phase of height 2^{kappa m}, which require checking the sparse-uniform dichotomy, the TT* reductions, and the final D-type estimates. This is not a presentation issue; the stationary off-diagonal component is indispensable for the main result.","section":"Section 4.2.2, Theorem 4.1"},{"comment":"The treatment of the case k<0, which corresponds to the component BHC^Delta_- in the decomposition of Section 2.2, is only an outline. Theorem 7.1 is stated with 'we only provide a sketch of the proof' and the argument splits into two parts with references to 'bootstrapping', [24, Theorem 4.3], and [24, Proposition 4.9], but the actual estimates for the uniform component in the regime -m/2 <= k <= 0 are not given. Since Main Theorem 1.1 covers all k in Z, the boundedness of BHC^Delta_- is part of the claimed result and cannot be relegated to an appendix sketch without a complete proof.","section":"Section 7 (Appendix)"},{"comment":"Proposition 3.4 controls the light component V^{tilde p,L}_{m,k} and is essential for the constancy-propagation step in Case II (0 <= k <= m/2). The proof ends with a description of the diagonal and off-diagonal terms D^2_diag and D^2_off and states 'We leave the further details to the interested reader.' In particular, the claimed decays 2^{(delta+4mu-epsilon0/4)k} and the use of the curvature in the mapping q |-> sqrt(2^{m/2}u+q^2)/q are not demonstrated. Because this proposition is a central step in the proof of the main smoothing estimate, the details cannot be omitted.","section":"Section 3.5, Proposition 3.4"},{"comment":"The treatment of the uniform component in the regime m/2 <= k <= m (Case I) is described as following the argument of Proposition 3.4 'with the obvious adaptations', and the final estimate for the term tilde D^2 is said to follow 'using similar reasonings to the ones employed for treating (3.21)'. This case is needed to cover the full range of k and m in Theorem 3.1, and the omitted details include several nontrivial changes of variables and the estimate of the analogue of the term D in (3.20)-(3.22). The reader cannot verify that the claimed decay 2^{-tilde epsilon1 m} in (3.37) is obtained without a full argument.","section":"Section 3.7"}],"minor_comments":[{"comment":"The phrase 'throughout the reminder of the paper' should be 'throughout the remainder of the paper'.","section":"Section 2"},{"comment":"The definition of BHC^{notDelta,NS} says it corresponds to 'the non-stationary phase regime, i.e., when the first item above is not satisfied'; this is ambiguous because the previous list has items (a), (b), (c), and 'the first item' could be read as only (a). The intended meaning is that none of (a)-(c) holds.","section":"Section 2.2.2"},{"comment":"In the display defining the sparse and uniform index sets, the notation |I^{ik+m}_r| and the normalization with 3I^{0,ik+m}_r are not defined explicitly before use; a remark on the convention for intervals and their dilations would improve readability.","section":"Section 3.2"},{"comment":"Reference [1] appears to list the authors in an unusual order ('Lars Becker, van Floris Doorn...'); please verify the spelling and ordering of the author names.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The main issue is that the theorem as stated goes beyond what is proved. If the model-case proof is correct, the paper could be made acceptable by either (i) supplying a rigorous reduction from arbitrary a, alpha to the model case, including the negative-exponent and unordered-exponent regimes, or (ii) restating Main Theorem 1.1 for the model parameters and clearly marking the general-parameter case as an open conjecture. The latter option would be a substantial weakening but would preserve the core contribution. I would also suggest that the editors ask for full proofs of Theorems 4.1 and 6.1 and of the k<0 appendix before acceptance, since these are load-bearing and currently deferred."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe real content of this paper is a boundedness theorem for the Bilinear Hilbert–Carleson operator in the model case a=(1,-1,1), α=(1,2,3): sup_λ |p.v. ∫ f1(x-t) f2(x+t²) e^{iλt³} dt/t| maps L^{p1} × L^{p2} into L^r for 1/p1 + 1/p2 = 1/r, 1 < p1,p2 < ∞, r > 1/2. That is new, and it is a substantial result. The machinery is the Rank II LGC framework from [24] with a new constancy-propagation step for the linearizing phase; the sparse/uniform dichotomy and the treatments of the non-stationary and low-oscillatory components are detailed. The authors say plainly that Proposition 1.18 ([24, Theorem 4.3]) is the main smoothing input for k ≥ m, with the new material handling m/2 ≤ k ≤ m and 0 ≤ k ≤ m/2. That is heavy reliance on a long unpublished preprint by two of the authors, but it is not circular.\n\nThe main soft spot is the one in the stress-test note, and it is load-bearing. Section 2 fixes a=(1,-1,1), α=(1,2,3) for the rest of the paper, yet Main Theorem 1.1 claims boundedness for all nonzero a_j and all pairwise distinct nonzero real α_j. The only justification is the sentence in Section 1.3 that the methods handle any nonresonant choice \"with minimal effort\". That is an assertion, not a reduction, and it is not a reparametrization: a change of variables sending a_j φ(s)^{α_j} to c_j s^j forces α = c(1,2,3), so triples such as (1,2,-1) or (-1,-2,-3) are not equivalent to the model. The proof also uses the ordering |t|³ ≤ t² ≤ |t| for |t| ≤ 1 in Section 3.1 and the stationary-phase classification of Section 2.2 for the specific phase; negative or unordered exponents break both, and negative exponents change the singularity at t = 0. As written, the theorem in its stated generality is not established, even though the model-case statement is plausible and probably correct.\n\nSecondary gaps of the same kind: Theorem 4.1 and Theorem 6.1 are stated with details left to the reader, the proof of the central new Proposition 3.4 ends with the D² estimates delegated to [24], and the k < 0 case in the appendix is an outline. None of this looks like an error; it is incompleteness. The word \"maximal\" for the range also deserves a sentence identifying the r = 1/2 barrier (setting λ = 0 leaves a curved bilinear Hilbert transform along (t, t²)).\n\nSpecialists in time-frequency analysis will want to read this, and it deserves a serious referee. My recommendation: send to peer review, instructing that the reductions be carried out or the theorem restated for the proved case, and the deferred proofs completed.","headline":"New and substantial boundedness result for the model curve (t, t^2, t^3), but the full claim for all pairwise distinct real alpha is not proved: the reduction to the model case is asserted, and several key proofs are deferred.","tokens_in":78209,"tokens_out":10994,"would_cite":true,"duration_ms":103246,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["42B20","42B25"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that the bilinear Hilbert–Carleson operator along monomial curves with pairwise distinct exponents is bounded from $L^{p_1}\\times L^{p_2}$ to $L^r$ for every Hölder pair with $1/2<r<\\infty$, up to endpoints.","keywords":["Bilinear Hilbert-Carleson operator","curved trilinear Hilbert transform","Rank II LGC method","constancy propagation","time-frequency analysis","wave packet decomposition","quasi-Banach range","non-resonant"],"falsifier":"Check the promised transfer by deriving the stationary-phase decomposition for the curve $(t,t^2,t^4)$ or $(t,t^2,t^{1/2})$. If the discriminant and root symmetry in Section 2.1 — $t_0=-2^k\\eta/(3\\lambda)$ and $|t_1-t_0|=|t_2-t_0|$ — fails for exponents other than $(1,2,3)$, then the division into diagonal, stationary off-diagonal, and non-stationary components no longer follows, and the generality claim in Main Theorem 1.1 would need a separate proof.","tokens_in":77069,"feed_emoji":"","tokens_out":10567,"duration_ms":98885,"temperature":0.7,"pith_summary":"This paper establishes the full expected boundedness range, up to endpoints, for the bilinear Hilbert–Carleson operator along monomial curves whose three exponents are pairwise distinct. For any nonzero coefficients and any distinct nonzero exponents $\\alpha_1,\\alpha_2,\\alpha_3$, the operator $BHC_{[\\vec a,\\vec\\alpha]}(f_1,f_2)(x)=\\sup_{\\lambda\\in\\mathbb{R}}|\\,p.v.\\int f_1(x-a_1t^{\\alpha_1})f_2(x-a_2t^{\\alpha_2})e^{i\\lambda a_3t^{\\alpha_3}}\\,dt/t|$ maps $L^{p_1}\\times L^{p_2}$ into $L^r$ whenever $1/p_1+1/p_2=1/r$ and $1/2<r<\\infty$. The proof is written only for the model curve $(t,t^2,t^3)$ with coefficients $(1,-1,1)$; the theorem asserts that every other pairwise-distinct choice follows from the same machinery with minimal effort. A sympathetic reader should take the paper as a full proof of the model case and as a documented claim of generality whose main theorem is conditional on that transfer. The result matters because it completes the purely non-zero curvature branch of a hierarchy in which the zero-curvature bilinear Hilbert–Carleson operator and trilinear Hilbert transform remain open.","feed_headline":"Full boundedness range for curved bilinear Carleson operators","feed_subtitle":"A correlative time-frequency method controls the λ-supremum on distinct-exponent curves for every r > 1/2.","key_machinery":"The load-bearing object is the Rank II LGC method, a correlative time-frequency/wave-packet discretization that tracks simultaneous Fourier-mode interactions of the two input functions and of the linearizing phase $\\lambda(x)$. The decisive mechanism inside it is the constancy propagation procedure: a bootstrap that lengthens the scale on which $\\lambda(x)$ is constant from $2^{-m-2k}$ to $2^{-m-k}$, applied after a sparse-uniform dichotomy separates spatially concentrated from uniformly distributed pieces of $f_1$ and $f_2$. Once the phase is constant enough, the maximal joint Fourier coefficient $J_{m,k}$ is controlled either by the LGC smoothing estimate from the companion curved trilinear Hilbert transform work (when $k\\ge m/2$) or by a Rank-I wave-packet model with a time-frequency correlation set (when $0\\le k\\le m/2$). A refined frequency localization for the third input in the stationary off-diagonal component makes the sums over $k$ and $m$ converge.","core_discovery":"Main Theorem 1.1 states that for $\\vec a=(a_1,a_2,a_3)$ and $\\vec\\alpha=(\\alpha_1,\\alpha_2,\\alpha_3)$ with each $\\alpha_j$ nonzero and pairwise distinct, the operator $BHC_{[\\vec a,\\vec\\alpha]}(f_1,f_2)(x)=\\sup_{\\lambda\\in\\mathbb{R}}|\\,p.v.\\int f_1(x-a_1t^{\\alpha_1})f_2(x-a_2t^{\\alpha_2})e^{i\\lambda a_3t^{\\alpha_3}}\\,dt/t|$ obeys $\\|BHC_{[\\vec a,\\vec\\alpha]}(f_1,f_2)\\|_{L^r}\\lesssim_{\\vec a,\\vec\\alpha,r,p_1,p_2}\\|f_1\\|_{L^{p_1}}\\|f_2\\|_{L^{p_2}}$ for every Hölder range $1/p_1+1/p_2=1/r$ with $1<p_1,p_2<\\infty$ and $1/2<r<\\infty$. The proof reduces the analysis to the representative case $\\vec a=(1,-1,1)$, $\\vec\\alpha=(1,2,3)$, that is, $BHC(f_1,f_2)(x)=\\sup_{\\lambda}|\\,p.v.\\int f_1(x-t)f_2(x+t^2)e^{i\\lambda t^3}\\,dt/t|$. In this model the operator is split into low-oscillatory, diagonal, and stationary/non-stationary off-diagonal components. The main work is a local $L^2$ smoothing estimate for the diagonal component: the maximal joint Fourier coefficient $J_{m,k}(f_1,f_2)$ along the moment curve satisfies $\\|J_{m,k}(f_1,f_2)\\|_{L^2(I_k^r)}\\lesssim 2^{-\\epsilon\\min\\{2k,m\\}}\\|f_1\\|_{L^2(3I_k^r)}\\|f_2\\|_{L^2(3I_k^r)}$, and this decay is extended to the full quasi-Banach range by multilinear interpolation with restricted weak-type bounds.","pith_inferences":["The asserted transfer from $(1,2,3)$ to arbitrary pairwise distinct exponents is the point a skeptical reader should test first: for non-integer or negative $\\alpha_j$, the phase derivative and stationary-point analysis change shape, and the paper does not display the reduction.","A concrete testable extension is whether the proof survives the third exponent not equaling a sum or difference of the first two; the model $(t,t^2,t^3)$ has special algebraic relations, while the versatility claim says all distinct triples are equally easy.","The threshold $r>1/2$ mirrors the sharp range of the flat bilinear Hilbert transform, suggesting the curved non-resonant setting is no harder than the flat case in terms of output exponents; this is an interpretation, not a result of the paper."],"forward_implications":["If Main Theorem 1.1 holds, the purely non-resonant bilinear Hilbert–Carleson operator is bounded on the maximal expected quasi-Banach range $r>1/2$ with no restriction beyond the Hölder relation and pairwise distinct exponents.","The result completes the $n=3$ non-zero curvature hierarchy: together with the companion curved trilinear Hilbert transform bounds, both the maximal Carleson-type operator and the curved trilinear Hilbert transform are now settled in the purely non-resonant regime.","The constancy propagation procedure provides a template for operators whose linearized discretized models are not absolutely summable within a single scale, by exploiting hidden cancellation in joint Fourier coefficients."],"supporting_citations":[{"why":"Introduces the Rank II LGC method and supplies the constancy-propagation smoothing estimate (its Theorem 4.3, quoted as Proposition 1.18) that controls the diagonal component.","marker":"[24]"},{"why":"Proved the hybrid quadratic non-resonant case, which sets the benchmark and the comparison for the purely non-resonant result.","marker":"[3]"},{"why":"Provides the Rank I LGC method and the γ-Carleson operator estimates used for stationary and low-oscillatory components.","marker":"[42]"},{"why":"Gives the unified treatment of curved operators along hybrid curves and the shifted square-function estimates used for the non-stationary off-diagonal component.","marker":"[16]"},{"why":"Established the bilinear Hilbert transform bounds whose range $r>2/3$ is the classical benchmark that the curved result extends toward $r>1/2$.","marker":"[30]"}],"fun_headline_variants":["Maximal range for curved bilinear Hilbert-Carleson","Curved Carleson: optimal bounds for all r>1/2","Distinct-exponent curves: full Carleson boundedness","Bilinear Hilbert-Carleson on curves: end-point range"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof is written only for the curve $(t,t^2,t^3)$ with coefficients $(1,-1,1)$; the theorem's full generality rests on the asserted but unproven claim that every other pairwise-distinct choice of nonzero exponents and coefficients can be treated by the same method with minimal effort.","fun_headline_variants_meta":{"raw":{"variants":["Maximal range for curved bilinear Hilbert-Carleson","Curved Carleson: optimal bounds for all r>1/2","Distinct-exponent curves: full Carleson boundedness","Bilinear Hilbert-Carleson on curves: end-point range"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00066,"raw_usage":{"total_tokens":3226,"prompt_tokens":1359,"completion_tokens":1867,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":975,"completion_tokens_details":{"reasoning_tokens":1792}},"tokens_in":975,"tokens_out":1867,"duration_ms":13204,"temperature":1.0,"reasoning_tokens":1792,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T19:47:27.746225+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Check the promised transfer by deriving the stationary-phase decomposition for the curve $(t,t^2,t^4)$ or $(t,t^2,t^{1/2})$. If the discriminant and root symmetry in Section 2.1 — $t_0=-2^k\\eta/(3\\lambda)$ and $|t_1-t_0|=|t_2-t_0|$ — fails for exponents other than $(1,2,3)$, then the division into diagonal, stationary off-diagonal, and non-stationary components no longer follows, and the generality claim in Main Theorem 1.1 would need a separate proof.","supporting_citations":[{"cited_title":"The non-resonant bilinear Hilbert-Carleson operator.Adv","cited_arxiv_id":null,"evidence_quote":"Proved the hybrid quadratic non-resonant case, which sets the benchmark and the comparison for the purely non-resonant result."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Rank I LGC method and the γ-Carleson operator estimates used for stationary and low-oscillatory components."},{"cited_title":"Lp estimates on the bilinear Hilbert transform for 2 < p <∞","cited_arxiv_id":null,"evidence_quote":"Established the bilinear Hilbert transform bounds whose range $r>2/3$ is the classical benchmark that the curved result extends toward $r>1/2$."}],"review_version":1}