{"id":"6cdfe61c-67e0-4167-9bd6-9d99e3f18b51","arxiv_id":"2505.22136","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"The additive Lebesgue-type measure is spectral exactly when t1+t2 is a nonzero integer or t1-t2 is a nonzero integer; the Plus space has no exponential orthogonal basis.","lead":"A mathematics paper proves a complete rule for when a plus-shaped combination of two intervals has an exponential orthogonal basis, and it settles the previously open Plus space case. The same paper gives sharp bounds connecting frame constants to the typical gap sizes in a spectrum.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 4.5's limit inference is invalid: sinπλ'_1 ± sinπλ'_2 = 0 does not imply λ'_1 = λ'_2; this is the load-bearing step in the necessity proof of Theorem 1.7.","rationale":"The central claim is Theorem 1.7, and its necessity direction rests on Proposition 4.5. Sufficiency (Proposition 4.8) is an explicit construction and is the mathematically secure half. The reader's weakest-assumption identification, Lemma 4.1, is a real weakness because the proof is figure-based, but it is a bounded numerical lemma that can likely be made rigorous by elementary estimates on sinc. The more load-bearing defect is the displayed limit inference in Proposition 4.5: sinπλ'_1 ± sinπλ'_2 = 0 simply does not imply λ'_1 = λ'_2, and that implication is what eliminates all non-diagonal possibilities. I checked the natural repair: comparing the signed equation at two indices in a constant-sign subsequence yields λ'_1 = λ'_2 by cross-multiplication, so the intended argument is plausible. Because the gap is concrete and repairable but the printed proof is not valid at that point, the reader's CONDITIONAL verdict remains appropriate; I do not see grounds to upgrade to ACCEPT or to move to REJECT. Hence the verdict is unchanged, and my agreement with the reader's choice of weakest assumption is only partial.","tokens_in":14017,"tokens_out":22351,"duration_ms":225456,"concrete_test":"Re-derive the last block of Proposition 4.5 with two indices: choose an infinite subsequence of lattice points (n_k,n_k) on which the sign in (4.5) is constant, and write the zero-set equation for n=n_i and n=n_j with i≠j, without taking any limit. If sinπλ'_1 and sinπλ'_2 are nonzero, cross-multiplication forces (λ'_2−n_i)/(λ'_1−n_i) = (λ'_2−n_j)/(λ'_1−n_j), hence λ'_1 = λ'_2. Then handle the cases sinπλ'_1 = 0 or sinπλ'_2 = 0 explicitly. If the same-sign subsequence cannot be chosen or exceptional solutions such as λ'_1 + λ'_2 ∈ Z survive, Proposition 4.5 needs a new argument; if the comparison works, the gap is a fixable omission rather than a fatal flaw.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Proposition 4.5, after obtaining infinitely many lattice points (n_k,n_k) in Λ, the proof takes a non-lattice point (λ'_1,λ'_2), passes to the limit k→∞ in the zero-set equation, and obtains sinπλ'_1 ± sinπλ'_2 = 0. It then asserts 'which implies that λ'_1 = λ'_2'. This implication is false: sinπa + sinπb = 0 holds for a−b ∈ 2Z+1 or a+b ∈ 2Z, while sinπa − sinπb = 0 holds for a−b ∈ Z or a+b ∈ 2Z+1; none of these forces a=b. This step is exactly what reduces Λ to the diagonal and leads to the arithmetic conclusion t1−t2 ∈ Z or t1+t2+1 ∈ Z, so the necessity direction of Theorem 1.7 is not established as written. The step is repairable by comparing the same signed zero-set equation at two indices with constant sign instead of taking a limit; however, as printed the inference is invalid. The reader's Lemma 4.1 concern is genuine but secondary: Lemma 4.1 is a bounded numerical claim that can likely be proved by monotonicity of sinc, whereas the Proposition 4.5 inference is false as stated.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces essential minimal and maximal spectral gaps for frame spectra and derives quantitative relations between frame bounds and these gaps: Theorem 1.1 gives gmin(Λ)gmax(Λ) ≤ C²π²/A under a |µ̂(ξ)| ≤ C|ξ|⁻¹ decay assumption, Theorem 1.3 gives a limsup bound for A·gmin(Λ_A) in terms of the L² norm of the density, and Theorem 1.5 gives upper and lower bounds on the maximal gap for frames of L[0,1]. In the second half, the paper applies these estimates to additive Lebesgue-type measures ρ_{t1,t2}, proving that the Plus space L²(ρ_{-1/2}) has no exponential orthonormal basis (Proposition 4.3) and classifying spectrality of ρ_{t1,t2} (Theorem 1.7): ρ_{t1,t2} is spectral if and only if t1+t2 ∈ Z\\{-1} or t1−t2 ∈ Z\\{0}. The sufficiency direction is proved by explicit construction of spectra verified through the Jorgensen–Pedersen criterion; the necessity direction proceeds through frame bounds, Lemma 4.1, and Proposition 4.5.","tokens_in":14331,"tokens_out":25667,"duration_ms":235043,"significance":"If the gaps in the proofs are repaired, the paper would resolve an open question posed by Lai–Liu–Prince on the spectrality of Plus spaces and complete the classification of additive Lebesgue-type measures, extending recent results of Ai–Lu–Zhou and Kolountzakis–Wu. The quantitative frame-gap estimates are of independent interest and the explicit spectrum constructions in Proposition 4.8 are a clear strength. The paper is also honest about the role of external results, and the main classification is not circular: it does not assume the target theorem.","major_comments":[{"comment":"The limit step after Eq. (4.5) is invalid. From the zero-set equation the author obtains sinπλ'_1 ± sinπλ'_2 = 0 and then writes 'which implies that λ'_1 = λ'_2'. This implication is false: sinπa + sinπb = 0 holds when a−b ∈ 2Z+1 or a+b ∈ 2Z, and sinπa − sinπb = 0 holds when a−b ∈ Z or a+b ∈ 2Z+1; none of these conditions forces a=b. This inference is exactly what reduces Λ to the diagonal and leads to the arithmetic conclusion t1−t2 ∈ Z, so the necessity direction of Theorem 1.7 is not established as written. The step is repairable by comparing the signed zero-set equation at two lattice points with the same sign instead of taking a limit, but as printed the assertion is false.","section":"Proposition 4.5, case λ1=λ2"},{"comment":"The proof of Lemma 4.1 is incomplete: it relies on the phrase 'As can be seen from the figure' and on the unproved numerical bound |f(x0)| ≤ 0.22 for the minimal positive solution of tan x0 = x0. No analytic derivation is given for the claimed dichotomy λ1 = λ2 or λ1 + λ2 = 0 under |λ1| ≤ 0.8. Since this lemma is load-bearing for Proposition 4.3 and again for Proposition 4.5, the tangent-line/figure argument must be replaced by a complete analytic proof.","section":"Lemma 4.1"},{"comment":"The proof of the claim F(x) ≤ F(1/2) contains an incorrect displayed estimate. For x ≥ 0.65 the paper asserts F(x) ≤ x(1/x + 1/(x+1)²) + x∫_{x+1}^∞ t⁻² dt = 1 + 1/x − 1/(x+1)², but the natural bound from the preceding expression is 1/x + x/(x+1)² + x/(x+1); the displayed equality is false. The subsequent Taylor-series estimate for x ∈ [1/2, 0.65] is asserted through an unmotivated chain of inequalities that appears to contain algebraic errors and is not verifiable as written. Since this claim is exactly what yields gmin(Λ)gmax(Λ) ≤ C²π²/A, Theorem 1.1 is not proved as printed.","section":"Theorem 1.1, proof of the claim on F(x)"},{"comment":"The periodicity conclusion is imported without proof: 'Then, similar to the arguments of Lemma 5.4 of [15], τ1(Λ) is a periodic set with a period in Z.' This is a nontrivial transfer and is not established in the manuscript. In addition, the lower bound |g_k(τ_j(Λ))| ≥ min(1/(2|t1−t2|), 1/(2|t1+t2+1|)) does not follow from T(g_k(τ1),g_k(τ2)) ∈ Z\\{0}: for example, when t1=t2 the condition is |g1−g2| ≥ 1/|2t1+1|, which allows g1 to be arbitrarily small. Lemma 4.4 is needed for Proposition 4.5, so this gap must be filled.","section":"Lemma 4.4"},{"comment":"Even if one grants Λ ⊂ {(x,x)}, the step 'This means T(λ'_1,λ'_2) ∈ 2Z+1 when (λ'_1,λ'_2) ∈ Λ−Λ with λ'_1 ∉ Z. Then τ1(Λ) ⊂ Z ∪ (λ1+Z)' is not justified. The oddness condition gives 2(λ'−μ)(t1−t2) ∈ 2Z+1 for two points λ', μ of τ1(Λ), which does not by itself imply that every non-integer element of τ1(Λ) lies in λ1+Z. This conclusion is then used to derive t1−t2 ∈ Z, so the gap is load-bearing for Theorem 1.7.","section":"Proposition 4.5, passage after the diagonal reduction"}],"minor_comments":[{"comment":"The frame inequalities in the statements of Theorem 1.5, Theorem 1.3, and Proposition 2.1 are missing squares on the summands; as printed they are not the inequalities used in the proofs.","section":"Theorem 1.5 and Theorem 1.3"},{"comment":"The interval is written as '[0.1]' instead of '[0,1]' in the statement of Theorem 1.5 and in Proposition 2.1.","section":"Theorem 1.5"},{"comment":"There are typos: 'τ1(Λ1)' should be 'τ1(Λ)', and the denominator '2t′+1' appears to be a typo for '2t2+1'.","section":"Lemma 4.4"},{"comment":"Figure 1 is referenced but no figure is included, and 'Figue' is a typo; the numerical constant 0.22 should be derived analytically rather than read from a graph.","section":"Lemma 4.1 and Figure 1"},{"comment":"The definition of C mixes a summand with a summed quantity; the notation should be clarified so that the displayed formula for Σ|ρ̂(ξ+λ)|² is unambiguous.","section":"Proposition 4.8"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses a significant open question and the overall strategy is plausible, but several load-bearing steps in the necessity proof are either false as written or only justified by reference to a figure or to an external lemma. I recommend major revision rather than rejection, since the invalid limit inference in Proposition 4.5 appears repairable and the remaining issues, while substantial, are local to specific proofs."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You asked about Lu's arXiv:2505.22136. The short version: the paper has two genuinely new quantitative estimates and a clean sufficiency construction, but the necessity proof of the main classification theorem has a logical gap that the stress-test note correctly identifies.\n\nThe new ideas are real. The essential min/max spectral gap notion in Theorem 1.1, the decay-rate estimate in Theorem 1.3 with its sharp example in Proposition 2.1, and the maximal gap bound in Theorem 1.5 together give a fresh quantitative picture of Landau's density theorem. The sufficiency direction of Theorem 1.7, Proposition 4.8, is a nice piece of work: the choice of spectrum and the Poisson-sum cancellation via Lemma 4.7 are convincing. The paper is also honest about borrowing from Lai-Liu-Prince and Kolountzakis-Wu; self-citation is not an issue here.\n\nThe soft spots are real and need attention before this paper can stand. In Proposition 4.5, the author finds infinitely many lattice points (n_k, n_k) in the spectrum, takes a non-lattice point (λ'_1, λ'_2), passes to the limit in the zero-set equation, and obtains sin πλ'_1 ± sin πλ'_2 = 0. The assertion that this implies λ'_1 = λ'_2 is false: sin πa = -sin πb only gives a+b ∈ 2Z, and sin πa = sin πb only gives a-b ∈ Z. Neither forces equality. This step is exactly what reduces Λ to the diagonal and leads to the arithmetic conclusion t_1 - t_2 ∈ Z or t_1 + t_2 + 1 ∈ Z, so the necessity direction of Theorem 1.7 is not established as written. The step may be repairable by fixing the sign across a sequence of n_k, but that repair is not in the paper.\n\nMinor but genuine issues: Lemma 4.1's proof is a figure and tangent-line heuristic, not a complete analytic argument; the 0.8 threshold deserves a real proof. In Theorem 1.1 the Taylor-expansion comparison F(x) ≤ F(1/2) is hard to follow and the displayed inequality has a garbled step. Lemma 4.4 imports a periodicity conclusion from Kolountzakis-Wu with only \"similar to\" justification, which is too thin for a load-bearing step.\n\nWho is this for? Specialists in spectral measures and Fourier frames. It deserves a serious referee because the claimed classification is sharp, the sufficiency is solid, and the problem is important. But the current version should not be accepted: a referee should insist on a corrected necessity proof and a proper write-up of Lemma 4.1.\n\nRecommendation: send to a good harmonic analysis journal with major revision required.","headline":"The classification theorem is plausible and the sufficiency construction is elegant, but the necessity proof contains a false sine implication that is load-bearing for the main result.","tokens_in":14865,"tokens_out":2099,"would_cite":false,"duration_ms":23142,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["28A80","42C05"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that the additive Lebesgue-type measure $\\rho_{t_1,t_2}$ is spectral exactly when $t_1+t_2$ is an integer other than $-1$ or $t_1-t_2$ is a nonzero integer.","keywords":["frame-spectral measure","essential minimal spectral gap","essential maximal spectral gap","spectral measure","Plus space","additive measure of Lebesgue type","Landau's theorem","Hurwitz zeta function"],"falsifier":"Run a bounded numerical search over $t_1,t_2$ and $\\lambda_1 \\in [-0.8,0.8]$ for a solution of the zero-set equation $e^{\\pi i(\\lambda_1(2t_1+1) - \\lambda_2(2t_2+1))} \\sin(\\pi\\lambda_1)/(\\pi\\lambda_1) + \\sin(\\pi\\lambda_2)/(\\pi\\lambda_2) = 0$ with $\\lambda_1 \\neq \\lambda_2$ and $\\lambda_1 + \\lambda_2 \\neq 0$; finding one would disprove Lemma 4.1 and thereby break Propositions 4.3 and 4.5.","tokens_in":4270,"feed_emoji":"📐","tokens_out":6233,"duration_ms":144493,"temperature":0.7,"pith_summary":"This paper links the two-sided frame bounds of a measure to the distribution of gaps in its frame spectra and uses that link to settle an open question about orthogonal bases for a family of two-line measures. It introduces the essential minimal and maximal spectral gaps $g_{\\min}(\\Lambda)$ and $g_{\\max}(\\Lambda)$, and proves that a Fourier decay condition $|\\hat{\\mu}(\\xi)| \\leq C |\\xi|^{-1}$ forces $g_{\\min}(\\Lambda) g_{\\max}(\\Lambda) \\leq C^2 \\pi^2 / A$. For Lebesgue measure on an interval, the paper obtains the two-sided estimate $4/(\\pi^2 B) \\leq g_{\\max}(\\Lambda) \\leq \\pi^2 B/A + 2$. Applying these gap bounds to the additive Lebesgue-type measure $\\rho_{t_1,t_2} = \\tfrac12(\\mathcal{L}_{[t_1,t_1+1]} \\times \\delta_0 + \\delta_0 \\times \\mathcal{L}_{[t_2,t_2+1]})$, the paper proves that it is a spectral measure exactly when $t_1+t_2 \\in \\mathbb{Z} \\setminus \\{-1\\}$ or $t_1-t_2 \\in \\mathbb{Z} \\setminus \\{0\\}$. In particular, the Plus space $L^2(\\rho_{-1/2})$ has no exponential orthogonal basis, answering the open question.","feed_headline":"Plus space has no exponential orthogonal basis","feed_subtitle":"Frame-gap bounds settle which additive Lebesgue-type measures are spectral.","key_machinery":"The essential minimal and maximal spectral gaps, defined by $g_{\\min}(\\Lambda) = \\inf\\{c \\geq 0 : g_k(\\Lambda) < c \\text{ for infinitely many } k\\}$ and $g_{\\max}(\\Lambda) = \\sup\\{c \\geq 0 : g_k(\\Lambda) > c \\text{ for infinitely many } k\\}$, carry the argument by converting the classical density theorem for frame spectra into gap constraints that depend on frame bounds. The quantitative engine is the Fourier lower-frame inequality: testing $e^{2\\pi i \\xi x}$ gives $A \\leq C^2 \\sum 1/|\\lambda_k - \\xi|^2$, and bounding this sum by a Hurwitz zeta function yields the product bound of Theorem 1.1. For the spectrality application, the load-bearing geometric input is Lemma 4.1: if $(\\lambda_1,\\lambda_2)$ lies in the zero set of the Fourier transform and $|\\lambda_1| \\leq 0.8$, then $\\lambda_1 = \\lambda_2$ or $\\lambda_1 + \\lambda_2 = 0$. The paper also uses a projection lemma, which says that the coordinates of a spectrum project to tight frame-spectra of the unit interval with frame bound $2$, reducing two-dimensional orthogonality to one-dimensional frame-gap constraints.","core_discovery":"The central claim is the classification of additive Lebesgue-type measures: $\\rho_{t_1,t_2}$ is spectral if and only if $t_1+t_2 \\in \\mathbb{Z} \\setminus \\{-1\\}$ or $t_1-t_2 \\in \\mathbb{Z} \\setminus \\{0\\}$. The forward direction shows existence by exhibiting explicit spectra: when $t_1-t_2$ is a nonzero integer one takes the integer lattice together with a shifted copy, and similarly when $t_1+t_2$ is an integer other than $-1$, using the sum rule for Fourier transforms to verify the orthogonal-basis criterion. The reverse direction rules out every other parameter pair. It first shows that any spectrum projects to tight frame-spectra of the unit interval with frame bound $2$, so the new gap estimates force arbitrarily small gaps in the projection. A geometric lemma on the zero set of the Fourier transform then pushes those small gaps onto the diagonal or anti-diagonal, and a finite-local-complexity periodicity argument finishes by pinning the parameters to the two arithmetic conditions. A direct consequence is that the Plus space $L^2(\\rho_{-1/2})$ admits no exponential orthogonal basis.","pith_inferences":["The $0.8$ threshold in Lemma 4.1 is probably not sharp; the same diagonal/anti-diagonal rigidity may hold for a larger radius, which would simplify verification of the geometric step without changing the main theorem.","Theorem 1.3 suggests a general uncertainty trade-off for absolutely continuous measures: the minimal gap of frame spectra with lower bound $A$ decays at least as fast as $A^{-1} \\|g\\|_2^2$, and this principle may extend to higher dimensions with the appropriate density.","The projection argument may transfer to other finite unions of orthogonal subspaces: spectrality could be probed by projecting a putative spectrum onto each component and checking one-dimensional frame-gap constraints.","A testable extension is to replace the interval length $1$ by general length $L$ in Corollary 1.4 and check whether the sharp constant becomes $L$."],"forward_implications":["For any frame-spectrum of a measure with $|\\hat{\\mu}(\\xi)| \\leq C |\\xi|^{-1}$, the product of the essential minimal and maximal gaps is at most $C^2 \\pi^2 / A$, so stronger lower frame bounds force the spectrum to develop both arbitrarily small and larger-than-average spacings.","For Lebesgue measure on $[0,1]$, every $A,B$ frame-spectrum has essential maximal gap trapped between $4/(\\pi^2 B)$ and $\\pi^2 B/A + 2$.","A tight frame-spectrum of $L^2([0,1])$ has all consecutive spectral gaps bounded above by $\\pi^2 + 2$.","The classification of additive Lebesgue-type measures is complete: such a measure is spectral exactly when $t_1+t_2 \\in \\mathbb{Z} \\setminus \\{-1\\}$ or $t_1-t_2 \\in \\mathbb{Z} \\setminus \\{0\\}$.","The Plus space $L^2(\\rho_{-1/2})$ has no exponential orthogonal basis."],"supporting_citations":[{"why":"Landau's density theorem supplies the classical lower-density constraint on frame spectra that motivates the essential-gap quantities.","marker":"[18]"},{"why":"Formulates the additive-measure spectrality question and supplies the zero-set characterization of the Fourier transform used throughout Section 4.","marker":"[17]"},{"why":"Handles the symmetric case $t \\neq -1/2$ that Theorem 1.7 extends to the two-parameter family.","marker":"[1]"},{"why":"Establishes non-spectrality for irrational parameters and contributes the finite-local-complexity periodicity argument adapted in Lemma 4.4.","marker":"[15]"},{"why":"Supplies the necessary-and-sufficient orthogonal-basis criterion used to certify the explicit spectra in Proposition 4.8.","marker":"[12]"}],"fun_headline_variants":["Plus space lacks exponential basis, gap bounds prove it","Gap classifications settle additive spectral measures","No exponential basis for Plus space: new gap theorem","Frame-gap estimates resolve spectrality of Plus space"],"cache_read_input_tokens":17024,"weakest_assumption_plain":"The load-bearing premise is Lemma 4.1, which says that if $(\\lambda_1,\\lambda_2)$ is a zero of the Fourier transform of $\\rho_{t_1,t_2}$ and $|\\lambda_1| \\leq 0.8$, then $\\lambda_1 = \\lambda_2$ or $\\lambda_1 + \\lambda_2 = 0$; the lemma's proof is a tangent-line and figure comparison, not a complete analytic derivation, and the later propositions rely on it directly.","fun_headline_variants_meta":{"raw":{"variants":["Plus space lacks exponential basis, gap bounds prove it","Gap classifications settle additive spectral measures","No exponential basis for Plus space: new gap theorem","Frame-gap estimates resolve spectrality of Plus space"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000231,"raw_usage":{"total_tokens":1447,"prompt_tokens":870,"completion_tokens":577,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":486,"completion_tokens_details":{"reasoning_tokens":517}},"tokens_in":486,"tokens_out":577,"duration_ms":6596,"temperature":1.0,"reasoning_tokens":517,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T13:24:45.753119+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run a bounded numerical search over $t_1,t_2$ and $\\lambda_1 \\in [-0.8,0.8]$ for a solution of the zero-set equation $e^{\\pi i(\\lambda_1(2t_1+1) - \\lambda_2(2t_2+1))} \\sin(\\pi\\lambda_1)/(\\pi\\lambda_1) + \\sin(\\pi\\lambda_2)/(\\pi\\lambda_2) = 0$ with $\\lambda_1 \\neq \\lambda_2$ and $\\lambda_1 + \\lambda_2 \\neq 0$; finding one would disprove Lemma 4.1 and thereby break Propositions 4.3 and 4.5.","supporting_citations":[{"cited_title":"Landau, Necessary density conditions for sampling and interpolation of certain entire functions, Acta Math","cited_arxiv_id":null,"evidence_quote":"Landau's density theorem supplies the classical lower-density constraint on frame spectra that motivates the essential-gap quantities."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Formulates the additive-measure spectrality question and supplies the zero-set characterization of the Fourier transform used throughout Section 4."},{"cited_title":"Ai, Z.-Y","cited_arxiv_id":null,"evidence_quote":"Handles the symmetric case $t \\neq -1/2$ that Theorem 1.7 extends to the two-parameter family."},{"cited_title":"Spectrality of a measure consisting of two line segments","cited_arxiv_id":"2501.11367","evidence_quote":"Establishes non-spectrality for irrational parameters and contributes the finite-local-complexity periodicity argument adapted in Lemma 4.4."},{"cited_title":"Jorgensen, S","cited_arxiv_id":null,"evidence_quote":"Supplies the necessary-and-sufficient orthogonal-basis criterion used to certify the explicit spectra in Proposition 4.8."}],"review_version":1}