{"id":"1d77b718-e951-449f-8db5-028e8af8e0db","arxiv_id":"1908.09548","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For every symmetric quasi-Banach space inside the Lorentz space Λ log, the paper constructs and proves optimal (smallest) symmetric quasi-Banach range spaces for the Calderón operator, the Hilbert transform, and the triangular truncation operator.","lead":"This paper identifies the smallest symmetric spaces into which the Calderón operator, the Hilbert transform, and the triangular truncation operator always map a given symmetric quasi-Banach input space. It extends classical results of Boyd, Gohberg-Krein, and Arazy to quasi-Banach and noncommutative settings, with applications to operator Lipschitz functions and commutator estimates.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 23's parity error breaks the lower-bound proof of Theorem 34: the constructed c vanishes on the parity that Hd requires, so the displayed lower estimate cannot hold as written.","rationale":"The paper's central contribution is the explicit identification of the optimal (smallest) symmetric quasi-Banach range space for the Calderón operator and, via it, for the Hilbert transform and triangular truncation operator. The upper-bound inclusions are plausibly supported by Theorem 14, and the construction of F is natural. The load-bearing weak point identified by the reader is genuine and internal: Lemma 23 contains an outright parity inconsistency. The definition (5.12) gives c nonzero only where k is even and nonpositive, while (5.13) demands k ≡ n+1 (mod 2); for even n this means all contributing k are odd, where c vanishes. The claimed lower bound for even n is therefore impossible. Because this lemma is the engine of Theorem 21 and hence of the minimality half of Theorem 34, the published proof does not currently establish that the range space F(H) is optimal. The same defect is not needed for Theorem 33, whose lower bound is delegated to a classical cited result, so the reader's formulation slightly overstates the scope of this particular gap. The issue is repairable: switching the lower estimate to odd n and then passing to decreasing rearrangements likely restores the comparison; a sample computation for a = e0 shows the odd-indexed values of Hdc are comparable to Sd e0 after rearrangement. Thus a conditional verdict is appropriate rather than rejection, and the reader's CONDITIONAL assessment should stand unchanged.","tokens_in":32079,"tokens_out":12306,"duration_ms":124523,"concrete_test":"Recompute the key estimate of Lemma 23 with the parity-corrected target: define c(0)=a(0), c(-2j)=a(j) for j≥1, and c(k)=0 otherwise; then for odd n=2m+1 compute (Hdc)(2m+1) = (2/π)∑_{j≥0} a(j)/(2m+1+2j). Verify whether there is a universal C such that |(Hdc)(2m+1)| ≥ C^{-1} Sdµ(a)(2m+1) for all m, and whether the decreasing rearrangement of this odd-indexed sequence is comparable to Sdµ(a). Test the comparison explicitly for a = e0 and a(k)=1/(k+1). If the comparison fails in either case, the stated construction cannot support the lower bound of Theorem 34 without a substantive new argument.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The optimality (minimality) half of Theorem 34 is proved through Theorem 21, which in turn depends on Lemma 23. In Lemma 23, the sequence c defined in (5.12) is nonzero only at nonpositive even indices and is zero at all positive indices. However, Hd, defined in (5.4) and used in (5.13), sums only over k satisfying k ≡ n+1 (mod 2). For every even n ≥ 0, the permitted k are odd, and all odd k carry c(k) = 0; hence (Hdc)(n) = 0 for every even n. This directly contradicts the displayed estimate |(Hdc)(n)| ≥ (1/2π) Sdµ(n,a) for n ≥ 0, n ≡ 0 (mod 2). The inequality can only make sense for n odd, and even then the constants and the final rearrangement step need to be checked. As written, the proof of Sdµ(a) ≤ cabs µ(Hdc) is not valid, so Theorem 21 and the optimality claim of Theorem 34 are not established by the stated argument. Theorem 33's lower bound uses [3, Proposition III.4.10] rather than Lemma 23, so the reader's statement slightly overstates the reach of this defect; its direct load is on Theorem 34 and on the applications depending on it (Theorem 37). The defect is localized and appears repairable by correcting the parity and redrawing the comparison, but the published proof is internally inconsistent.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the optimal symmetric quasi-Banach range spaces for the Calderón operator S on (0,∞), for the noncommutative Hilbert transform 1⊗H on spaces over M⊗L∞(R), and for the triangular truncation T on ideals of compact operators. The central construction is F = {x : μ(x) ≤ Sμ(y) for some y ∈ E}, equipped with a quasi-norm defined by an infimum of ‖y‖E. The authors prove an abstract domination theorem (Theorem 14) showing that a self-adjoint contraction with suitable L^p and weak-L^1 bounds is dominated by S on Λlog(M), and they use this to establish the upper-bound halves of the optimal-range theorems. The lower-bound halves are proved via the classical Bennett–Sharpley result for the Hilbert transform and via a new noncommutative analogue (Theorem 21) for T. Section 8 applies these results to double operator integrals and commutator/Lipschitz estimates. The stated goal is to resolve Problems 1 and 2 in full generality, removing the Fatou-norm assumption used in earlier Boyd-type results.","tokens_in":32383,"tokens_out":5797,"duration_ms":54915,"significance":"If correct, the paper would solve two longstanding problems in the optimal-range theory of symmetric quasi-Banach spaces and would provide a unified noncommutative framework for the Hilbert transform and the triangular truncation. The construction of F is explicit, parameter-free, and defined directly from the operator S, with no fitted constants; the upper-bound arguments in Lemmas 18 and 19 and Theorem 14 are careful and mostly complete. The optimality claim for the triangular truncation, however, rests on Lemma 23, which contains a genuine parity mismatch that invalidates the proof as written. This defect is localized and appears repairable, but it blocks the minimality half of Theorem 34 and the applications that depend on it. The lower bound for the Hilbert transform in Theorem 33 is not affected, since it uses the external classical result [3, Proposition III.4.10].","major_comments":[{"comment":"The sequence c defined in (5.12) is nonzero only at nonpositive even integers and is zero at all positive integers and at all odd integers. For n ≥ 0 even, the parity condition in (5.13) (k ≡ n+1 mod 2) selects only odd k, for which c(k)=0 by (5.12). Hence (Hdc)(n)=0 for every even n ≥ 0, so the displayed lower bound |(Hdc)(n)| ≥ (1/2π) Sdμ(n,a) cannot hold for these n. Since the subsequent rearrangement step relies on this estimate, the proof of Lemma 23 and therefore of Theorem 21 is not valid as written.","section":"§5, Lemma 23 (Eqs. (5.12)–(5.13))"},{"comment":"The optimality (minimality) direction of Theorem 34 is proved by invoking Theorem 21, whose proof depends on Lemma 23. With Lemma 23 uncorrected, the conclusion F(H) ⊂ G(H) for an arbitrary symmetric quasi-Banach ideal G with T(E(H)) ⊂ G(H) is not established. The applications in Theorem 37 that cite Theorem 34 are consequently unsupported as written. This defect does not affect the lower bound in Theorem 33, which cites [3, Proposition III.4.10] rather than Lemma 23; however, Theorem 34 is one of the headline results.","section":"§7, Theorem 34 and §8, Theorem 37"}],"minor_comments":[{"comment":"There is a sign inconsistency: (5.4) defines (Hda)(n) with denominator (k−n), whereas (5.11) and Lemma 22 use denominator (n−k) with the same parity condition. Because all later estimates are taken in absolute value, this does not alter the inequalities, but the two definitions should be reconciled.","section":"§5, Eq. (5.4) vs. Eq. (5.11)"},{"comment":"The sentence 'A ≈ B means A ≲ B and A /greaterorsimilarB' contains an unrendered symbol; it should read 'A ≳ B'.","section":"§3, after Theorem 14"},{"comment":"In the optimality part of the proof, after showing Sμ(x) ∈ G(0,∞) for x ∈ E(0,∞), the text says 'Hence, F(0,∞) ⊂ G(0,∞)' without explicitly invoking the definition of F. The inference is correct but should be spelled out for clarity.","section":"§6, Proof of Theorem 26"}],"recommendation":"major_revision","confidential_remarks":"The paper is substantial and the overall strategy is coherent, but the Lemma 23 parity error is a genuine correctness gap in a load-bearing lemma. I recommend major revision rather than rejection, because the defect appears localized and repairable: the construction of c can be adjusted to put mass on the parity class required by Hd, or Hd can be redefined to use the opposite parity. The authors should also reconcile the sign between (5.4) and (5.11). The upper-bound halves and Theorem 33 are in much better shape and should be preserved."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a substantial paper and the main construction is real. The idea of defining the optimal range space F through domination by the Calderón operator S is clean and genuinely new, and it correctly removes the Fatou/Banach assumptions that Boyd, Gohberg–Krein, and Arazy all needed. The upper-bound machinery in Theorem 14 is careful and mostly convincing, and the applications to operator Lipschitz functions and commutator estimates are valuable. If the proof gaps I name below are fixed, this will be a significant paper.\n\nThe soft spots are localized but real. Lemma 23 is the load-bearing step for the lower bound in Theorem 34. As printed, (5.12) defines c(k) to be nonzero only for even nonpositive k, with c(k)=0 for all odd k and all positive k. But the operator Hd in (5.4) sums only over k ≡ n+1 (mod 2). For even n, that means k is odd, so every term vanishes and (Hdc)(n)=0. This directly contradicts the displayed estimate |(Hdc)(n)| ≥ (1/2π) Sdμ(n,a) for even n. The repair is almost certainly to change the parity condition to k ≡ n (mod 2), or equivalently to place the values on the opposite parity; with that change the displayed sum matches the intended argument. But as it stands, Theorem 21 and the optimality half of Theorem 34 are not established by the written proof. The sign inconsistency between (5.4) and Lemma 22 is immaterial for absolute values but should be cleaned up.\n\nThere is also a smaller overclaim in Theorem 14(i): the statement covers all of Λ log, but the proof through Lemma 19 only reaches the smaller space Λψ. The later use in Theorem 14(ii) is saved because the element A2 is bounded and does lie in Λψ, so the applications still go through, but the theorem as stated is stronger than what is proved.\n\nThe optimality proof for the Hilbert transform in Theorem 33 rests on the classical Bennett–Sharpley lower bound, not on Lemma 23, so that part is on solid ground. The citation pattern looks honest and the new results are clearly separated from old ones.\n\nWho should read this: specialists in symmetric spaces, noncommutative integration, and operator ideals. It deserves a serious referee — the defects are concrete but localized and almost certainly repairable, and the framework is worth engaging with carefully. I would send it to review with a request to fix Lemma 23 and the statement of Theorem 14(i).","headline":"Genuinely new framework for optimal range of the Hilbert transform and triangular truncation in symmetric quasi-Banach spaces, but the discrete lower-bound proof has a concrete parity error in Lemma 23 that needs repair before Theorem 34 is proved as written.","tokens_in":32907,"tokens_out":4329,"would_cite":true,"duration_ms":43768,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["46E30","47B10","46L51","46L52","44A15","47L20","47C15"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper identifies the optimal (smallest) symmetric quasi-Banach range space for the Calderón operator, the Hilbert transform, and the triangular truncation operator, for every symmetric quasi-Banach domain inside the Lorentz space…","keywords":["symmetric quasi-Banach spaces","Calderón operator","Hilbert transform","triangular truncation","optimal range","operator Lipschitz functions","commutator estimates","Lorentz spaces"],"falsifier":"Take $a$ with $\\mu(k,a)=1/(k+1)$ and let $c$ be as in (5.12). For every even $n$, $(H_d c)(n)=0$ because the parity condition only allows odd $k$, where $c$ vanishes; the claimed formula replacing the sum by $\\sum_{k\\ge0}\\mu(k,a)/(n+2k)$ is inconsistent with (5.12)–(5.13). Computing $S_d\\mu(a)(n)\\sim \\log n/n$ and comparing it with the decreasing rearrangement of the odd-index values of $H_d c$ would settle whether any absolute constant can make the domination hold.","tokens_in":31888,"feed_emoji":"🎯","tokens_out":12984,"duration_ms":110684,"temperature":0.7,"pith_summary":"The paper settles a long-standing question: given a symmetric quasi-Banach space $E$ of functions or sequences inside the Lorentz space $\\Lambda_{\\log}$, what is the smallest symmetric quasi-Banach space that receives the Calderón operator $S$, the Hilbert transform, and the triangular truncation operator? The answer is the same space $F$ in all three cases, defined by the condition that the rearrangement of the output is dominated by $S$ applied to the rearrangement of some element of $E$. This extends the classical reduction in [7] from the Banach-space setting to the full quasi-Banach setting, and it sharpens the classical weak-$L^1$ result: for $E = L^1$ the optimal range of the Hilbert transform is the closure of bounded functions in $L^{1,\\infty}$.","feed_headline":"Smallest range for Calderón operator, Hilbert transform identified","feed_subtitle":"For every symmetric quasi-Banach space, the optimal receptacle is built from the Calderón operator on rearrangements.","key_machinery":"The Calderón operator $S$, defined by $(Sx)(t) = \\frac{1}{t}\\int_0^t x(s)\\,ds + \\int_t^\\infty \\frac{x(s)}{s}\\,ds$, is the carrying mechanism. Theorem 14 converts quantitative $L^p$ bounds of an operator $T$ into a universal domination $\\mu(T(A)) \\preceq c_{abs} S\\mu(A)$; this is what makes $S$ the universal upper bound. Lemma 23 supplies the matching lower bound via the discrete operator $H_d$ and a rearrangement-preserving construction of a sequence $c$ from $a$, proving that triangular truncation is strong enough to force $F$ to be the smallest possible range.","core_discovery":"The central discovery is that for every symmetric quasi-Banach space $E \\subset \\Lambda_{\\log}$, the space $F = \\{x : \\mu(x) \\le S\\mu(y) \\text{ for some } y \\in E\\}$, equipped with the natural infimum norm, is itself a symmetric quasi-Banach space and is the optimal symmetric quasi-Banach range for the Calderón operator $S$. The same $F$ serves as the optimal range for the non-commutative Hilbert transform $1\\otimes H$ acting on $E(M \\bar{\\otimes} L^\\infty(\\mathbb{R}))$ when $M$ is atomless and semifinite, and for the triangular truncation operator $T$ acting on the ideal $E(H)$. The upper bound is proved through an abstract domination theorem: any self-adjoint contraction on $L^2(M)$ whose norm on $L^p(M)$ grows at most like $1/(p-1)$ satisfies $\\mu(T(A)) \\preceq c_{abs} S\\mu(A)$ in the submajorization sense. The lower bound, proving that no smaller symmetric quasi-Banach space can work, is built from a discrete Fourier construction showing that triangular truncation can already reproduce the Calderón operator on rearrangements.","pith_inferences":["Editorial inference: the same formula for $F$ may describe optimal range spaces for any operator dominated by $S$ in rearrangement, including Riesz projections and the absolute value map, giving a uniform template for rearrangement-invariant optimal range problems.","Editorial inference: replacing the triangular truncation lower bound with a direct construction for the Hilbert transform would isolate whether the parity issue in Lemma 23 is essential or an artifact; such a construction would make the proof robust and testable.","Editorial inference: should the Lemma 23 lower bound fail for some $E$, the minimality claim would still hold for symmetric spaces with Fatou norm via the classical argument; the quasi-Banach extension would then require a different lower-bound mechanism."],"forward_implications":["For $E=L^1$, the optimal range of the Hilbert transform is $(L^{1,\\infty})_0$, the closure of bounded functions in $L^{1,\\infty}$, refining the classical weak-$L^1$ theorem.","For $E=L^1(H)$, the optimal range of the triangular truncation operator is the weak trace ideal $L^{1,\\infty}(H)$, confirming the classical prediction for triangular truncation on the trace class.","For every Lipschitz function $f$ and self-adjoint operators $A,B$, the double operator integral maps $E(H)$ into $F(H)$ and satisfies $\\|[f(A),B]\\|_{F(H)} \\le c\\,\\|f'\\|_\\infty \\|[A,B]\\|_{E(H)}$; the same holds for $f(X)-f(Y)$ with $\\|X-Y\\|_{E(H)}$.","A noncommutative logarithmic integrability theorem follows: if $\\|\\mu(A)\\log^+(\\mu(A))\\|_{L^1}<\\infty$, then $T(A)\\in L^1(M)$ with the corresponding norm bound, covering both the Hilbert transform and triangular truncation."],"supporting_citations":[{"why":"Supplies the classical reduction that boundedness of the Hilbert transform is equivalent to boundedness of the Calderón operator for symmetric Banach spaces, the baseline this paper extends.","marker":"[7]"},{"why":"Provides the maximal domain $\\Lambda_{\\log}$ for the Hilbert transform and the rearrangement inequalities used throughout the upper-bound proof.","marker":"[3]"},{"why":"Establishes the $L^p$ growth condition for the non-commutative Hilbert transform, so that Theorem 14 applies to $1\\otimes H$.","marker":"[39]"},{"why":"Contains the classical predictions for the range of triangular truncation on the trace class, which are confirmed and sharpened as optimal range statements.","marker":"[21]"},{"why":"Provides the weak type estimate for the triangular truncation operator used in Theorem 11 and interpolation tools for the abstract domination theorem.","marker":"[16]"},{"why":"Supplies the double operator integral bounds that allow the $S$-domination to be transferred to Lipschitz functions and commutators in Theorem 37.","marker":"[11]"},{"why":"Supplies the completeness criterion for quasi-normed symmetric operator spaces used to show $F(H)$ is a quasi-Banach space.","marker":"[41]"}],"fun_headline_variants":["Optimal range for Calderón operator and Hilbert transform found","Sharpest range for Calderón, Hilbert, and truncation operators","Symmetry yields optimal range for Calderón and Hilbert transforms","One quasi-Banach range serves Calderón, Hilbert, and truncation","Calderón's best range also covers Hilbert and triangular truncation"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof that $F$ is the smallest possible range depends on Lemma 23, whose construction of the sequence $c$ assigns nonzero values only to even indices, while the parity condition $k \\equiv n+1 \\pmod{2}$ in the definition of $H_d$ excludes those indices when $n$ is even; as written, the displayed lower estimate for $(H_d c)(n)$ does not follow from the stated definitions.","fun_headline_variants_meta":{"raw":{"variants":["Optimal range for Calderón operator and Hilbert transform found","Sharpest range for Calderón, Hilbert, and truncation operators","Symmetry yields optimal range for Calderón and Hilbert transforms","One quasi-Banach range serves Calderón, Hilbert, and truncation","Calderón's best range also covers Hilbert and triangular truncation"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000302,"raw_usage":{"total_tokens":1690,"prompt_tokens":843,"completion_tokens":847,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":459,"completion_tokens_details":{"reasoning_tokens":758}},"tokens_in":459,"tokens_out":847,"duration_ms":8465,"temperature":1.0,"reasoning_tokens":758,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:11:18.593311+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take $a$ with $\\mu(k,a)=1/(k+1)$ and let $c$ be as in (5.12). For every even $n$, $(H_d c)(n)=0$ because the parity condition only allows odd $k$, where $c$ vanishes; the claimed formula replacing the sum by $\\sum_{k\\ge0}\\mu(k,a)/(n+2k)$ is inconsistent with (5.12)–(5.13). Computing $S_d\\mu(a)(n)\\sim \\log n/n$ and comparing it with the decreasing rearrangement of the odd-index values of $H_d c$ would settle whether any absolute constant can make the domination hold.","supporting_citations":[{"cited_title":"Boyd, The Hilbert transform on rearrangement-invariant spaces, Can","cited_arxiv_id":null,"evidence_quote":"Supplies the classical reduction that boundedness of the Hilbert transform is equivalent to boundedness of the Calderón operator for symmetric Banach spaces, the baseline this paper extends."},{"cited_title":"Bennett and R","cited_arxiv_id":null,"evidence_quote":"Provides the maximal domain $\\Lambda_{\\log}$ for the Hilbert transform and the rearrangement inequalities used throughout the upper-bound proof."},{"cited_title":"Randrianantoanina, Spectral subspaces and non-commutative Hilbert transform s, Colloq","cited_arxiv_id":null,"evidence_quote":"Establishes the $L^p$ growth condition for the non-commutative Hilbert transform, so that Theorem 14 applies to $1\\otimes H$."},{"cited_title":"Gohberg and M.G","cited_arxiv_id":null,"evidence_quote":"Contains the classical predictions for the range of triangular truncation on the trace class, which are confirmed and sharpened as optimal range statements."},{"cited_title":"Dodds, T.K","cited_arxiv_id":null,"evidence_quote":"Provides the weak type estimate for the triangular truncation operator used in Theorem 11 and interpolation tools for the abstract domination theorem."},{"cited_title":"Caspers, D","cited_arxiv_id":null,"evidence_quote":"Supplies the double operator integral bounds that allow the $S$-domination to be transferred to Lipschitz functions and commutators in Theorem 37."},{"cited_title":"Sukochev, Completeness of quasi-normed symmetric operator spaces , Indag","cited_arxiv_id":null,"evidence_quote":"Supplies the completeness criterion for quasi-normed symmetric operator spaces used to show $F(H)$ is a quasi-Banach space."}],"review_version":1}