{"id":"33baaf0a-9e6b-4581-8793-94c0f6a8ed73","arxiv_id":"2411.15564","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For flat symmetric spaces, L1-L2 dichotomy holds for all rank 1 cases except Cartan type AI, and holds at regular points in SU(2,q) and SU(3,q) but fails at certain singular points in SU(2,q).","lead":"This paper determines exactly when convolution powers of orbital measures on flat symmetric spaces are well-behaved: in rank 1, everything works except the space SL(2,R)/SO(2), and in SU(2,q) regular points work while certain singular points fail. It provides the first rank 2 examples of this failure, a new phenomenon in harmonic analysis.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Unproven extension of [12, Corollary 3] to rank 2 is load-bearing: Theorems 4.2(q=2) and 4.3 need µ_H^{*2} ∈ L1 for singular H to rule out singularity, and Lemma 2.3(2) supplies this only via an unproven parenthetical extension.","rationale":"The reader's weakest assumption is exactly this, and I agree. The remainder of the paper's analytic estimates (Sections 3 and 4) are internally consistent: the Plancherel weights match the root multiplicities, the Bessel asymptotics (4.5)-(4.8) are standard, and the domain decompositions cover the Weyl chambers. The paper honestly flags the open SU(2,2) type A case, but that is not the main risk because type D q=2 already provides a rank-2 failure if the L1 premise holds. The unproven parenthetical extension is the single point on which the novelty rests, so a conditional acceptance requiring the gap to be filled or proved is the right call. No change to the reader's conditional verdict is needed.","tokens_in":17876,"tokens_out":17936,"duration_ms":149237,"concrete_test":"Verify whether [12, Corollary 3] (or its proof) covers rank G/K = 2 for root systems C_2 and BC_2. If not, prove directly that for H = (x,x) in SU(2,2) and H = (x,0) in SU(2,3), the measure µ_H^{*2} is in L1(p) by evaluating its spherical transform via (4.1) and Bessel-function asymptotics; if neither the citation nor a direct proof supplies this, Theorems 4.2 (q=2) and 4.3 are not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Lemma 2.3(2) is the load-bearing bridge for the rank-2 failure claims. To conclude that L1-L2 dichotomy fails at a singular point H, the paper must know that µ_H^{*2} is not singular; the only argument is Lemma 2.3(2), which asserts µ_H^{*k(G)} ∈ L1(p) for every nonzero H when rank ≥ 2. The proof invokes [12, Corollary 3] with the parenthetical remark '(Though not explicitly stated in [12], the result is valid for rank G/K ≥ 2.)' and gives no derivation. The cited corollary is not shown to cover the rank-2 root systems C_2 and BC_2 used in SU(2,q) and SU(3,q). If the extension is false, µ_H^{*2} may be singular for the singular H of Theorems 4.2 and 4.3, so those theorems would not establish failure of the dichotomy and the paper's claimed first rank-2 examples would be unsupported. This is a gap in proof, not a disagreement with consensus.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the L1-L2 dichotomy for orbital measures on flat symmetric spaces G0/K = K ⋉ p / K, using the spherical transform and Bessel-function asymptotics. In rank 1 it proves that the dichotomy holds for every non-compact irreducible symmetric space except Cartan type AI. For the AIII spaces SU(2,q)/(S(U(2)×U(q))) it proves the dichotomy for regular points, proves that type D singular points satisfy the dichotomy when q>2 but fail when q=2, and proves that type A singular points fail for q≥3 with an explicit L2 threshold; the type A q=2 case is left open. The rank 3 regular case is also treated. The main new claimed contribution is the first rank-2 examples of failure of the L1-L2 dichotomy on flat symmetric spaces.","tokens_in":73,"tokens_out":40518,"duration_ms":475166,"significance":"If the proofs are correct, the paper gives a substantial advance: a complete rank-1 classification and the first rank-2 examples where the L1-L2 dichotomy fails, with quantitative thresholds. The paper is careful in stating the open type A q=2 case and in describing limitations of the method. The use of explicit spherical functions and detailed Bessel estimates is well suited to the problem, and the rank-1 part is convincing. However, two load-bearing points need repair before the rank-2 and rank-3 claims are supported: an unproved extension of a cited result in Lemma 2.3, and an algebraic inconsistency in the derivation of the main integrand (4.3). These are gaps in proof, not disagreements with consensus.","major_comments":[{"comment":"The passage from (4.1) to (4.3) is not algebraically correct. Substituting the displayed formula for ψ(λ,X) into |ψ(λ,X)|^{2k} and multiplying by δ(λ) gives the factor (λ_i^2−λ_j^2) in the denominator with exponent 2k−2, not in the numerator as written in (4.3); it also introduces an additional factor (λ_1...λ_p)^{2k(r+1)}. Hence every estimate in Section 4 is applied to an expression that is not the Plancherel integrand derived from the stated spherical function. This affects Theorems 4.1, 4.2, 4.3 and 4.5. The displayed formula for ψ should also be checked against the rank-1 formula (3.1): for p=1 the product ∏(x_kλ_k) gives ψ∝tλ J_{q−1}(tλ), which contradicts (3.1).","section":"Section 4.1, Eqs. (4.1)-(4.3)"},{"comment":"Lemma 2.3(2) asserts that for rank G/K ≥2, μ_H^{*k(G)} ∈ L1(p) for every nonzero H, and the proof invokes an extension of [12, Corollary 3] with the parenthetical remark that the result is valid for rank G/K ≥2 although not stated in [12]. This extension is load-bearing: Theorems 4.2 and 4.3 use μ_H^{*2} ∈ L1 for singular H in the root systems C_2 and BC_2. Without a proof or a precise citation verifying that the rank-2 root systems are covered, the nonsingularity of the second convolution power, and hence the failure claims at type D (q=2) and type A (q≥3), is not established.","section":"Lemma 2.3(2)"},{"comment":"Even accepting the integrand (4.3) as the intended one, the estimate on W1 does not prove the claimed convergence for k≥2 when q=2. The displayed bound gives φ ≤ C λ_1^{-((2r+1)(k−1)+2k−2)}, and for r=0, k=2 this is λ_1^{-3}; the polar-coordinate integral in (4.9) is then ∫ ρ dρ/ρ^3, which diverges. The authors state that (4.9) converges for k≥2, but the q=2 case is not covered by the displayed exponent. The proof needs an additional estimate that exploits the vanishing of the determinant when λ_1=λ_2.","section":"Theorem 4.1, estimate on W1"}],"minor_comments":[{"comment":"The displayed equality of the two expressions for the intervals I_n is not correct: the two products differ by a shift of π in the first factor. The subsequent lower-bound estimates appear to use the first expression, so the paragraph should be rewritten with consistent interval notation.","section":"Theorem 4.2, divergence intervals"},{"comment":"The value f_r(0) is written as 1/2^r, but from J_r(s)∼(s/2)^r/Γ(r+1) the correct value is 1/(2^r Γ(r+1)). This does not affect the stated integrability results, but the constant should be corrected.","section":"Section 4.1, definition of f_r"},{"comment":"The displayed estimate in (4.14) appears to have reciprocal exponents: the text says φ ≈ C' λ_1^{1+2r}/λ_1^{2(2k−2)} = C'/λ_1^{4k−5−2r}, which is not algebraically correct. The intended exponent should be stated explicitly so that the convergence condition can be checked.","section":"Theorem 4.3, Eq. (4.14)"},{"comment":"The reference [24] has a typo in the author's name ('Jodeph A. Wolf'); it should be 'Joseph A. Wolf'.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The novelty is potentially significant, but the proofs need substantial repair. The unproven rank-2 extension of [12] is acknowledged by the authors themselves, and the integrand inconsistency in (4.3) should be addressed before the paper can be accepted. It would be useful to ask the authors to supply a self-contained proof of Lemma 2.3(2) and to recompute the rank-2/3 estimates from the correct Plancherel integrand; the conclusions may survive, but the present manuscript does not fully support them."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a genuine first. No one has studied L1-L2 dichotomy for flat symmetric spaces before, and the paper gives a complete rank-1 answer plus the first rank-2 examples, including failures. I think the results are probably right, but there is one load-bearing gap that needs fixing.\n\nWhat's new and good: The setup is natural (orbital measures on the Euclidean space p via the semidirect product G0 = K ⋉ p), and the spherical transform approach works cleanly. Theorem 3.1 completes the rank-1 picture in parallel with the known compact and non-compact cases, and the AI exception (SL(2,R)/SO(2)) falls out of the Bessel asymptotics at ν = 0. The rank-2 results for SU(2,q) are the first of their kind: regular points satisfy the dichotomy, singular type D holds for q > 2 but fails for q = 2, and singular type A fails for q ≥ 3 with the sharp threshold k ≥ 3/4 + q/2. The proofs are detailed and the Bessel estimates are standard but carefully handled.\n\nThe soft spot is Lemma 2.3(2). It asserts that for rank ≥ 2, some fixed convolution power of any nonzero orbital measure lands in L1(p), citing [7, Corollary 7] improved by [12, Corollary 3]. The authors add the parenthetical that [12, Corollary 3] is valid for rank ≥ 2 even though not stated there. No derivation is given. This is load-bearing: the type D q = 2 failure and the type A q ≥ 3 failures require knowing that µ_H^{*2} is absolutely continuous so that it is not singular; without that, the dichotomy could hold vacuously at those points. The gap is real, but it is also small—it is a missing proof, not an example of circular reasoning. The rest of the paper does not depend on this in a way that would collapse; the rank-1 theorem and the regular-point theorems use only Lemma 2.3(1), which is solid.\n\nMinor open case: SU(2,2) type A is left undecided, and the authors say so in Remark 4.4. That is fine.\n\nMy reading: the citation pattern is honest—[1], [12], [14] are prior published results, not self-serving assumptions. The paper is worth refereeing, but the referee should insist that the authors supply a proof of the rank ≥ 2 part of Lemma 2.3(2) or restate it as a conjecture and adjust the theorems that rely on it. If that gap can be filled, this is a solid paper for the harmonic analysis community.","headline":"First study of L1-L2 dichotomy for flat symmetric spaces, with a clean rank-1 classification and the first rank-2 examples, but the rank-2 failure results rest on an unproven extension of a cited result that needs to be fixed.","tokens_in":18680,"tokens_out":2418,"would_cite":true,"duration_ms":19999,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["43A90","43A85","22E30"],"pacs":[],"model":"deepseek-v4-flash","headline":"Flat symmetric spaces satisfy the L1-L2 dichotomy in rank one except SL(2,R), and rank-two SU(2,q) spaces fail it at certain singular points.","keywords":["symmetric space","orbital measure","spherical function","L1-L2 dichotomy","flat symmetric space","Bessel functions","Plancherel formula","SU(p,q)"],"falsifier":"Check whether Lemma 2.3(2) is actually proved for rank $\\ge2$ by the cited result: if [12, Corollary 3] does not cover the needed root systems, then the absolute-continuity premise for the rank-2 and rank-3 dichotomy theorems collapses. Concretely, test a singular axis point in $\\mathrm{SU}(2,3)$: if $\\mu_H^{*2}$ is singular rather than absolutely continuous, the dichotomy would hold at that point, contradicting Theorem 4.3; if it is absolutely continuous, the failure claim reduces to the threshold calculation.","tokens_in":17674,"feed_emoji":"⚖️","tokens_out":9966,"duration_ms":76589,"temperature":0.7,"pith_summary":"This paper studies the $L^1$-$L^2$ dichotomy for flat (Euclidean) symmetric spaces: the property that every $K$-invariant orbital measure, under repeated convolution, is either singular or becomes square-integrable. The authors prove that every non-compact irreducible rank-one flat symmetric space satisfies the dichotomy except the Cartan type AI case, $\\mathrm{SL}(2,\\mathbb{R})_0/\\mathrm{SO}(2)$. For the rank-two and rank-three spaces $\\mathrm{SU}(p,q)_0/\\mathrm{S}(\\mathrm{U}(p)\\times\\mathrm{U}(q))$ with $p=2,3$, the dichotomy holds at regular points; at singular points of the rank-two family it holds for diagonal points when $q>2$, fails for diagonal points when $q=2$, and fails for axis points when $q\\ge 3$, with the sharp square-integrability threshold $k\\ge \\tfrac{3}{4}+\\tfrac{q}{2}$. These are the first rank-2 examples of failure of the dichotomy for any symmetric space. The argument reduces the question to the integrability of powers of explicit Bessel-function spherical transforms against the Plancherel measure.","feed_headline":"First rank-2 failures of L1-L2 dichotomy on flat symmetric spaces","feed_subtitle":"Rank-one spaces obey it except SL(2,R); SU(2,q) singular points fail for certain q.","key_machinery":"The central object is the orbital measure $\\mu_H$ on the flat symmetric space $\\mathfrak p\\cong G_0/K$, defined by averaging over the $K$-orbit of $H\\in\\mathfrak p$. Its spherical transform is the basic spherical function $\\psi_\\lambda(H)=\\int_K e^{iB(E_\\lambda,\\mathrm{Ad}(k)H)}\\,dk$, and the transform of a convolution product is the product of the transforms (Lemma 2.1). The Plancherel formula (Lemma 2.2) then identifies $\\|\\mu_H^{*k}\\|_{L^2(\\mathfrak p)}^2$ with an integral over the positive Weyl chamber of $|\\psi_\\lambda(H)|^{2k}$ against the weight $\\delta(\\lambda)=\\prod_{\\alpha\\in\\Phi^+}|\\alpha(E_\\lambda)|^{m_\\alpha}$. For rank one and for type AIII, the spherical functions are given explicitly by Ben Saïd–Ørsted in terms of Bessel functions: a single $J_\\nu$ in rank one, and a determinant of $J_{q-p}$ functions in the $\\mathrm{SU}(p,q)$ case. Asymptotic estimates for Bessel functions and their derivatives then determine, for each orbit, the smallest $k$ for which the Plancherel integral converges.","core_discovery":"On the paper's own terms, the central discovery is that flat symmetric spaces inherit the $L^1$-$L^2$ dichotomy from their rank-one compact and non-compact cousins, with the same single exception ($AI\\cong BDI(q=2)$), and that in rank two the dichotomy is controlled by the type of singularity. Specifically, for $G_0/K=\\mathrm{SU}(2,q)_0/\\mathrm{S}(\\mathrm{U}(2)\\times\\mathrm{U}(q))$, the dichotomy holds for regular orbital measures for all $q\\ge2$, holds for singular points of type D exactly when $q>2$, fails for type D when $q=2$, and fails for type A when $q\\ge3$; in the last case $\\mu_H^{*k}$ is square-integrable precisely when $k\\ge \\tfrac{3}{4}+\\tfrac{q}{2}$. For $p=3$, the regular-point dichotomy holds as well. If correct, these are the first rank-2 instances in which the dichotomy fails for any symmetric space.","pith_inferences":["The threshold $k\\ge \\tfrac{3}{4}+\\tfrac{q}{2}$ grows with $q$, suggesting that axis singularities become progressively harder to smooth as the ambient dimension grows; one testable extension is whether $\\mathrm{SU}(p,q)$ with $p\\ge3$ has similar thresholds with $p$-dependent constants.","The unresolved $q=2$ axis case is a natural sharp test: deciding the integrability of $\\int\\int |J_0(x\\lambda_1)-J_0(x\\lambda_2)|^4 \\lambda_1\\lambda_2/(\\lambda_1^2-\\lambda_2^2)^2\\,d\\lambda_1 d\\lambda_2$ would settle whether $\\mathrm{SU}(2,2)$ fails the dichotomy at both singular types or only at the diagonal one.","The same Bessel-function integrability method should apply to other flat symmetric spaces with explicit spherical functions, such as rank-two CII or BDI spaces, to test whether rank-2 failure is special to type AIII.","Because the absolute-continuity step for rank $\\ge2$ rests on a parenthetical extension of a cited result, a direct proof of $\\mu_H^{*2}\\in L^1(\\mathfrak p)$ for these spaces would make the dichotomy results independent of that extension."],"forward_implications":["For every non-compact irreducible rank-one flat symmetric space except $\\mathrm{SL}(2,\\mathbb{R})_0/\\mathrm{SO}(2)$, every nonzero orbital measure has a convolution square in $L^2(\\mathfrak p)$, so all higher powers are square-integrable as well.","In $\\mathrm{SU}(2,q)$ flat spaces, regular orbital measures satisfy the dichotomy for every $q\\ge2$, and diagonal singular points satisfy it for $q>2$.","For singular axis points in $\\mathrm{SU}(2,q)$ with $q\\ge3$, the dichotomy fails, and the minimal power for square integrability grows linearly: $k\\ge \\tfrac{3}{4}+\\tfrac{q}{2}$.","For $\\mathrm{SU}(3,q)$, regular points satisfy the dichotomy, extending the regular-point result to rank three."],"supporting_citations":[{"why":"Supplies the explicit Bessel-function formulas for spherical functions on flat symmetric spaces, used for rank one and type AIII.","marker":"[3]"},{"why":"The absolute-continuity result for convolutions of orbital measures in symmetric spaces; the paper extends it to rank ≥ 2 for its absolute-continuity premise.","marker":"[12]"},{"why":"Establishes absolute continuity of convolution powers of orbital measures on Riemannian symmetric spaces, used in Lemma 2.3.","marker":"[7]"},{"why":"Gives the transfer from group-level absolute continuity to flat-space orbital measures, used in Lemma 2.3.","marker":"[1]"},{"why":"Supplies the Plancherel formula for symmetric spaces used in Lemma 2.2.","marker":"[21]"},{"why":"Provides the root-system and multiplicity data for the rank-one classification and the general Lie algebra setup.","marker":"[20]"},{"why":"Supplies the restricted root system and multiplicity tables for the type AIII spaces SU(p,q).","marker":"[5]"}],"fun_headline_variants":["Rank-2 flat spaces break L1-L2 dichotomy at certain singular points","First L1-L2 failures on rank-2 flat symmetric spaces","SU(2,q) singular points selectively violate L1-L2 dichotomy","Regular points hold, specific singularities fail L1-L2 for SU(2,q)","L1-L2 dichotomy fails on rank-2 flat spaces: type matters"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof that the second convolution power of every nonzero orbital measure is absolutely continuous in rank $\\ge2$ rests on an improvement of a cited theorem that the authors themselves note is 'not explicitly stated' in the source, for the needed rank range.","fun_headline_variants_meta":{"raw":{"variants":["Rank-2 flat spaces break L1-L2 dichotomy at certain singular points","First L1-L2 failures on rank-2 flat symmetric spaces","SU(2,q) singular points selectively violate L1-L2 dichotomy","Regular points hold, specific singularities fail L1-L2 for SU(2,q)","L1-L2 dichotomy fails on rank-2 flat spaces: type matters"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000643,"raw_usage":{"total_tokens":2947,"prompt_tokens":927,"completion_tokens":2020,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":543,"completion_tokens_details":{"reasoning_tokens":1920}},"tokens_in":543,"tokens_out":2020,"duration_ms":16145,"temperature":1.0,"reasoning_tokens":1920,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T14:09:58.744729+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Check whether Lemma 2.3(2) is actually proved for rank $\\ge2$ by the cited result: if [12, Corollary 3] does not cover the needed root systems, then the absolute-continuity premise for the rank-2 and rank-3 dichotomy theorems collapses. Concretely, test a singular axis point in $\\mathrm{SU}(2,3)$: if $\\mu_H^{*2}$ is singular rather than absolutely continuous, the dichotomy would hold at that point, contradicting Theorem 4.3; if it is absolutely continuous, the failure claim reduces to the threshold calculation.","supporting_citations":[{"cited_title":"Analysis on flat symmetric space s, J","cited_arxiv_id":null,"evidence_quote":"Supplies the explicit Bessel-function formulas for spherical functions on flat symmetric spaces, used for rank one and type AIII."},{"cited_title":"The absolute continuit y of convolu- tions of orbital measures in symmetric spaces","cited_arxiv_id":null,"evidence_quote":"The absolute-continuity result for convolutions of orbital measures in symmetric spaces; the paper extends it to rank ≥ 2 for its absolute-continuity premise."},{"cited_title":"Absolute continuity of convo lutions of orbital measures on Riemannian symmetric spaces, J","cited_arxiv_id":null,"evidence_quote":"Establishes absolute continuity of convolution powers of orbital measures on Riemannian symmetric spaces, used in Lemma 2.3."},{"cited_title":"Convolution of o rbital mea- sures in symmetric spaces, Bull","cited_arxiv_id":null,"evidence_quote":"Gives the transfer from group-level absolute continuity to flat-space orbital measures, used in Lemma 2.3."},{"cited_title":"Groups and Geometric analysis, American M ath Soc., Mathematical Surveys and Monographs 83, 2002","cited_arxiv_id":null,"evidence_quote":"Supplies the Plancherel formula for symmetric spaces used in Lemma 2.2."},{"cited_title":"Differential geometry, Lie groups and sym metric spaces, Academic Press, New York, 1978","cited_arxiv_id":null,"evidence_quote":"Provides the root-system and multiplicity data for the rank-one classification and the general Lie algebra setup."},{"cited_title":"Random matrix theory and symme tric spaces, Physics Reports 394 (2004), 41–156","cited_arxiv_id":null,"evidence_quote":"Supplies the restricted root system and multiplicity tables for the type AIII spaces SU(p,q)."}],"review_version":1}