{"id":"4395e0ed-24d6-49a6-9886-68aa230fed42","arxiv_id":"1908.01634","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Generalized Lutwak-Petty and Leng-Lu projection and intersection inequalities are proved for Minkowski and radial Minkowski valuations generated by even, zonal measures.","lead":"This paper proves new sharp volume inequalities for a broad family of shapes defined through Minkowski valuations. It generalizes classical Lutwak-Petty and Leng-Lu inequalities and shows that Lutwak's original inequalities are the strongest in the family.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proof of Theorem 1.2 drops the nth power in the step after (5.5): the printed inequality is too weak to imply the asserted sharp ball maximum.","rationale":"The paper’s central new theorem, Theorem 1.2, is not established by the written proof because the step after (5.5) omits the nth power: as printed, the inequality is not strong enough to imply the claimed sharp maximum. This is a concrete, located gap in the internal argument, and it directly affects the main claim. I do not see an internal contradiction that would falsify the theorem itself, and the corrected inequality appears to restore the proof. The reliance on Theorem 1.1 from [23] is a normal dependency, though it means the equality cases are inherited from that paper; I did not find evidence that this dependency is circular. Because the flaw is real but repairable, the reader’s CONDITIONAL verdict remains appropriate, and no verdict adjustment is needed.","tokens_in":22625,"tokens_out":42790,"duration_ms":373385,"concrete_test":"Re-derive the chain for n=3 with K1=K2=Bn and µ normalized by µ(S^2)=1/2. The sharp corrected bound gives (1/4)^3 = [V3(ΓµB3)/V3(B3)]·V3(Φµ,∗B3)·V3(B3)^2, while the printed bound gives only 1/4 ≥ [ratio]·P; numerically P(B,B) = 1/(64·ratio), so the printed line would permit products up to 16 times the true ball maximum. Replace the displayed inequality by (1/(n+1))^n ≥ [Vn(ΓµBn)/Vn(Bn)]·∏Vn(Ki)·Vn(Φµ,∗(...)) and check that the equality analysis in the following sentence remains valid; if it does, the flaw is a typographical omission rather than a substantive mathematical error.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In the proof of Theorem 1.2 (Section 5), the authors set L = Φµ,∗(K1,...,Kn−1) in Theorem 5.1 and obtain (5.5): V(K1,...,Kn−1, ΓµΦµ,∗(...)) = 1/(n+1). By (2.7), V(...)^n ≥ Vn(K1)···Vn(Kn−1)·Vn(ΓµΦµ,∗(...)), so the correct consequence is (1/(n+1))^n ≥ Vn(K1)···Vn(Kn−1)·Vn(ΓµΦµ,∗(...)). Combining this with Theorem 5.2 gives (1/(n+1))^n ≥ [Vn(ΓµBn)/Vn(Bn)]·Vn(K1)···Vn(Kn−1)·Vn(Φµ,∗(...)). The printed line instead reads 1/(n+1) ≥ [Vn(ΓµBn)/Vn(Bn)]·Vn(K1)···Vn(Kn−1)·Vn(Φµ,∗(...)). Since n ≥ 3, the printed line is strictly weaker than the correct bound, so it cannot yield the sharp inequality Vn(Φµ,∗(...))·∏Vn(Ki) ≤ Vn(Φµ,∗Bn)·Vn(Bn)^{n−1}; the subsequent assertion that this is precisely the desired inequality does not follow. The gap is repairable by restoring the nth power, but as written the proof of the central theorem is incomplete.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper generalizes Lutwak's mixed projection volume inequalities and the Leng-Lu intersection-body inequalities to polarizations of Minkowski valuations generated by even, zonal measures on the sphere. The main results are Theorem 1.2, asserting that among K1,...,K_{n-1} the product Vn(Φ^{µ,*}(K1,...,K_{n-1}))Vn(K1)···Vn(K_{n-1}) is maximized by Euclidean balls, with equality characterized by homothetic ellipsoids in the discrete case; Theorem 1.3, a Busemann-Petty type centroid inequality for the generalized centroid bodies Γ^{µ}; Theorem 1.4, domination of polar projection bodies by affine quermassintegrals; and Theorem 1.5, the dual inequalities for new radial Minkowski valuations Ψ^{τ}_i, with a comparison to the classical intersection bodies I_i. The proofs use Hölder, Jensen, the Aleksandrov-Fenchel inequality, spherical convolution, and Radon transforms, and they rely on Theorem 1.1 of the authors' companion paper [23] and on polarization results from [59].","tokens_in":22942,"tokens_out":23398,"duration_ms":203748,"significance":"If the stated results hold, this is a substantial contribution to affine isoperimetric theory: it unifies the Lutwak-Petty inequalities, the Busemann intersection inequality, the Leng-Lu inequalities, and their recent Minkowski-valuation generalizations, and it identifies the classical projection-body inequalities as the strongest members of a large family. The paper also contains a new integral representation and explicit constants, and it connects the results to Lutwak's conjecture on affine quermassintegrals and to Grinberg's inequalities. The proofs are analytic and do not involve parameter fitting or definitional circularity; the main dependency on Theorem 1.1 of [23] is an external input rather than a reduction of the new inequalities to themselves. The paper is significant for convex geometry and geometric tomography, provided the two technical issues raised below are corrected.","major_comments":[{"comment":"The displayed inequality after (5.5) is missing the nth power. From (5.5) one has V(K1,...,K_{n-1}, Γ^{µ}Φ^{µ,*}(K1,...,K_{n-1})) = 1/(n+1), and applying (2.7) gives V(...)^n ≥ Vn(K1)···Vn(K_{n-1})Vn(Γ^{µ}Φ^{µ,*}(...)). The correct consequence is therefore (1/(n+1))^n ≥ Vn(K1)···Vn(K_{n-1})Vn(Γ^{µ}Φ^{µ,*}(...)). Combining this with Theorem 5.2 yields (1/(n+1))^n ≥ [Vn(Γ^{µ}Bn)/Vn(Bn)] Vn(K1)···Vn(K_{n-1})Vn(Φ^{µ,*}(...)). The printed line uses 1/(n+1) instead of its nth power; since n ≥ 3 this is strictly weaker and, together with (5.4), gives only Vn(Φ^{µ,*}(...))∏Vn(K_i) ≤ (n+1)^{n-1}Vn(Φ^{µ,*}Bn)Vn(Bn)^{n-1}, not the asserted sharp inequality. The proof is repairable by restoring the power, but as written it does not establish Theorem 1.2.","section":"Section 5, proof of Theorem 1.2, display after (5.5)"},{"comment":"The constant in the right-hand inequality of (1.8) is inconsistent with the proof and appears to be inverted. The derivation at the end of Section 5 gives Vn(I_iL) ≤ (κ_{n-1}^n κ_n / κ_i^n) ∫_{Gr_{n,i}} Vi(L∩E)^n dν_i(E) = (κ_{n-1}^n / κ_n^{n-1}) \\tilde{A}_{n-i}(L)^n. With the constant κ_n^{n-1}/κ_{n-1}^n as printed, the inequality is false already for L = Bn and i = n-1, where Vn(IBn) = κ_n κ_{n-1}^n but the printed right side equals κ_n^{2n-1}/κ_{n-1}^n. The statement of Theorem 1.5 should be corrected to match the factor derived in the proof.","section":"Section 4, Theorem 1.5, equation (1.8)"}],"minor_comments":[{"comment":"The reference 'Theorem (5.1)' should be 'Theorem 5.1'.","section":"Section 5, proof of Theorem 1.2"},{"comment":"The sentence 'replacing K by ϑK' should read 'replacing L by ϑL', since the equivariance is applied to the star body L.","section":"Section 5, proof of Theorem 5.6"},{"comment":"The phrase 'using (5.11, followed by integration' is missing a closing parenthesis; it should read 'using (5.11), followed by integration'.","section":"Section 5, proof of Theorem 1.5"},{"comment":"The sentence 'it is an open problem wether ...' contains a typo; 'wether' should be 'whether'.","section":"Page 24"},{"comment":"In the displayed definition of \\tilde{A}_{n-i}(L), the integration domain is written as Gr_{n,k}; it should be Gr_{n,i}.","section":"Section 2, definition of dual affine quermassintegrals"},{"comment":"The constant in (5.3) should be typeset unambiguously as (n+1)^n Vn(Γ^{µ}L) Vn(Φ^{µ,*}Bn) Vn(Bn)^{n-1}; the current rendering 'κn−1 n' is hard to read and could be mistaken for κ_{n-1}^n rather than κ_n^{n-1}.","section":"Section 5, equation (5.3)"}],"recommendation":"major_revision","confidential_remarks":"The paper is mathematically substantial and the central ideas are sound, but the two issues above are load-bearing: the missing nth power in the proof of Theorem 1.2 and the inverted constant in Theorem 1.5. Both are repairable, but the manuscript cannot be accepted until they are fixed. I also recommend that the authors carefully proofread all displayed constants and explicitly state in the introduction that Theorem 1.1 of [23] is fully proved there and that its equality conditions are used essentially."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Alex—quick take on Berg–Schuster, arXiv:1908.01634.\n\nThe paper does what it says: it extends Lutwak's mixed projection inequalities and the Leng–Lu intersection inequalities to Minkowski valuations generated by even, zonal measures, and to a dual family of radial Minkowski valuations. The new statements are real: Theorems 1.2, 1.3, 1.4, 1.5, 3.7, 5.1, 5.2, and 5.6 are not in the literature, and the integral representation in Lemma 3.2 is a genuinely useful device for comparing these bodies to the classical projection and centroid bodies. The proofs are mostly clean applications of Aleksandrov–Fenchel, Hölder, and Jensen, with equality conditions worked out carefully. I also like that the paper ties the new inequalities to Lutwak's conjecture on affine quermassintegrals and to Grinberg's inequalities.\n\nThe soft spot is in the proof of the central theorem. In Theorem 1.2, after setting L = Φ^{µ,*}(K1,...,K_{n−1}), the authors get (5.5): V(K1,...,K_{n−1}, Γµ Φ^{µ,*}(...)) = 1/(n+1). Combining this with the Aleksandrov–Fenchel inequality (2.7) requires raising both sides to the nth power, and then applying Theorem 5.2 gives\n\n(1/(n+1))^n ≥ [Vn(Γµ Bn)/Vn(Bn)] · Vn(K1)···Vn(K_{n−1}) · Vn(Φ^{µ,*}(...)).\n\nThe printed line instead has 1/(n+1) on the left, which is strictly weaker for n ≥ 3 and cannot yield the sharp inequality. The sentence claiming that the printed line is precisely the desired inequality is therefore not right. The gap is obviously repairable—restore the nth power and use (5.4)—but as written, the proof of the paper's centerpiece is incomplete. This is a typo-level defect, but it needs fixing.\n\nThe other caveat is dependency, not circularity. The paper uses Theorem 1.1 from the same authors' earlier work [23] as a black box, plus polarization from [59]. That is a legitimate input, and the new inequalities do not secretly reduce to those inputs. The citation pattern is fine. The paper also honestly flags the open problem of whether the Lp Busemann–Petty inequality is the strongest in its family.\n\nWho is this for: convex geometers in Brunn–Minkowski theory and valuations. It deserves a serious referee. My recommendation: send to peer review and have the referee check the exponent in (5.5) and ask the authors to correct it. With that fix, this is a solid contribution.","headline":"Solid generalizations of the Lutwak–Petty and Leng–Lu inequalities, but the proof of Theorem 1.2 has a dropped nth power after (5.5) that makes the printed argument too weak; it is a repairable typo, not a fatal flaw.","tokens_in":23503,"tokens_out":4463,"would_cite":true,"duration_ms":35308,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["52A20","52A39","52A40","44A12"],"pacs":[],"model":"deepseek-v4-flash","headline":"For even, zonal measures, balls maximize the mixed volume product defining generalized projection bodies, extending Lutwak's inequalities and their duals.","keywords":["Minkowski valuations","projection bodies","intersection bodies","mixed projection inequalities","zonal measures","affine quermassintegrals","Busemann–Petty centroid inequality","Leng–Lu inequalities"],"falsifier":"For n = 3 and an even, zonal measure µ that is neither discrete nor a scalar multiple of spherical Lebesgue measure, choose two convex bodies K_1, K_2 and numerically compute V_3(Φ_µ,*(K_1,K_2)) V_3(K_1) V_3(K_2); a value greater than the same product for Euclidean balls would disprove Theorem 1.2. Alternatively, for a star body L and a zonal τ, check whether V_n(Ψ_τ,i L)/V_n(L)^i exceeds V_n(I_i L)/V_n(L)^i, which would contradict Corollary 5.7 or Theorem 1.5.","tokens_in":22430,"feed_emoji":"📐","tokens_out":9450,"duration_ms":80671,"temperature":0.7,"pith_summary":"The paper extends two classical families of volume inequalities in convex geometry — Lutwak's mixed projection inequalities and the Leng–Lu intersection inequalities — to polarizations of Minkowski valuations generated by even, zonal measures on the unit sphere. The central assertion is that Euclidean balls maximize the relevant volume products, with equality characterized by homothetic ellipsoids in the discrete projection-body case. If true, this shows that the known affine isoperimetric inequalities are the strongest members of a whole family of Euclidean inequalities, and that the classical projection and intersection body results are sharp within these families. The dual half of the paper introduces radial Minkowski valuations that generalize intersection bodies and proves the analogous maximization, recovering the Busemann and Leng–Lu inequalities as special cases.","feed_headline":"Balls maximize a generalized Lutwak–Petty product","feed_subtitle":"Euclidean balls are extremal for a whole family of Minkowski valuations; classical projection inequalities win.","key_machinery":"The load-bearing object is the Minkowski valuation Φ_µ defined by h(Φ_µ K,u) = ∫_{∂K} h(Z_µ(u), ν_K(x)) $dH^{{n-1}}$(x), where Z_µ is the zonoid generated by an even, zonal measure µ; its polarization gives the mixed operators Φ_µ(K_1,...,K_{n-1}). On the dual side, the radial Minkowski valuations Ψ_τ are defined by ρ(Ψ_τ L, ·) = ρ(L,·)^{n-1} * Rτ, with R the spherical Radon transform. Lemmas 3.2 and 4.1 express Φ_µ,i and Ψ_τ,i as averages of the classical projection bodies Π_i and intersection bodies I_i over rotations; these integral representations, combined with Jensen and Hölder inequalities, carry the volume product estimates and identify the classical inequalities as the strongest in their families.","core_discovery":"The paper's main discovery is Theorem 1.2 and its dual Theorem 1.5. For every even, zonal measure µ on $S^{{n-1}}$, the mixed volume product V_n(Φ_µ,*(K_1,...,K_{n-1})) V_n(K_1)···V_n(K_{n-1}) is maximized by Euclidean balls; when µ is discrete, the maximizers are precisely homothetic ellipsoids, and the new radial valuations Ψ_τ,i satisfy the analogous inequality against intersection bodies and dual affine quermassintegrals. The proofs pass through a mixed-volume inequality (Theorem 5.1) and a generalized Busemann–Petty centroid inequality (Theorem 5.2), and through Jensen-type averaging arguments that single out the classical projection and intersection bodies as the strongest inequalities in the family.","pith_inferences":["Beyond the paper: the domination results (1.7) and (1.8) suggest that the classical projection and intersection bodies are not just special cases but the sharp members of whole one-parameter families; this could be tested numerically for intermediate zonal measures, where the inequalities predict a monotone volume product as µ interpolates between spherical Lebesgue measure and a discrete measure.","Beyond the paper: the same rotation-averaging technique used in Theorem 1.4 might extend to L_p and Orlicz settings beyond the zonal class, since Lemma 3.6 provides the needed averaging identity for Γ_µ,p and the paper already proves the L_p analogue for centroid bodies.","Beyond the paper: if Lutwak's affine quermassintegral conjecture (Conjecture 2.1) is ever proved, Theorem 1.4 will immediately upgrade every generalized Lutwak–Petty inequality to an affine invariant one, whereas the present paper already shows the reverse direction fails at the level of these families."],"forward_implications":["Corollary 5.3: for 1 ≤ i ≤ n−2, the volume product V_n(Φ_µ,i^* K) V_n(K)^i is maximized by Euclidean balls, interpolating between classical isoperimetric inequalities for quermassintegrals and the Lutwak–Petty projection inequalities.","Theorem 1.4: with µ(S^{n-1}) = 1/2, the polar volume of Φ_µ,i K is dominated by that of the classical projection body Π_i K, and both are dominated by the reciprocal of the affine quermassintegral A_{n-i}(K), so Lutwak's affine quermassintegral conjecture would imply these inequalities.","For every even, zonal τ with τ(S^{n-1}) = κ_{n-1}, the volume of the new radial valuation Ψ_τ,i L is at most V_n(I_i L), so the Busemann and Leng–Lu intersection inequalities dominate the whole family.","Theorem 3.7 and Corollary 5.5: an L_p analogue of the Busemann–Petty centroid inequality holds for generalized centroid bodies Γ_µ,p, and the polar volume product V_n(Γ_µ,p^* L) V_n(L) is maximized by balls, extending the L_p Busemann–Petty and Blaschke–Santaló results."],"supporting_citations":[{"why":"Supplies Theorem 1.1, the black-box volume product maximization, and Lemma 3.1, the averaging identity for Φ_µ over rotations.","marker":"[23]"},{"why":"Supplies the polarization formula (1.2) establishing existence of the mixed Minkowski valuations Φ_µ and the definition of generalized centroid bodies Γ_µ.","marker":"[59]"},{"why":"Lutwak's mixed projection inequalities, the discrete case that Theorem 1.2 generalizes.","marker":"[33]"},{"why":"Leng–Lu intersection inequalities of all orders, the results that Theorem 1.5 generalizes.","marker":"[27]"},{"why":"Provides the L_p affine isoperimetric inequalities and the equivalence technique used to prove Theorems 3.4, 3.7, and Corollary 5.5.","marker":"[43]"},{"why":"Grinberg's affine isoperimetric inequalities for dual affine quermassintegrals, used in the right-hand bound of (1.8).","marker":"[21]"},{"why":"Busemann's intersection inequality, the classical result recovered in the dual setting.","marker":"[7]"},{"why":"Introduces affine quermassintegrals and Conjecture 2.1, connecting Theorem 1.4 to a major open problem.","marker":"[36]"}],"fun_headline_variants":["Euclidean balls extremize generalized Lutwak–Petty inequalities","Balls maximize volume products for even zonal Minkowski valuations","Minkowski valuations: balls are extremal for generalized inequalities","Generalized projection inequalities: balls maximize mixed volumes","Euclidean balls extremize new volume inequalities for valuations"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proofs treat Theorem 1.1 from the authors' earlier article as a black box, so the new mixed inequalities inherit its validity; if that volume-product maximization fails for some even, zonal measure, the conclusions of Theorems 1.2 and 5.2 collapse.","fun_headline_variants_meta":{"raw":{"variants":["Euclidean balls extremize generalized Lutwak–Petty inequalities","Balls maximize volume products for even zonal Minkowski valuations","Minkowski valuations: balls are extremal for generalized inequalities","Generalized projection inequalities: balls maximize mixed volumes","Euclidean balls extremize new volume inequalities for valuations"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001501,"raw_usage":{"total_tokens":5966,"prompt_tokens":831,"completion_tokens":5135,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":447,"completion_tokens_details":{"reasoning_tokens":5051}},"tokens_in":447,"tokens_out":5135,"duration_ms":37473,"temperature":1.0,"reasoning_tokens":5051,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:07:15.354066+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For n = 3 and an even, zonal measure µ that is neither discrete nor a scalar multiple of spherical Lebesgue measure, choose two convex bodies K_1, K_2 and numerically compute V_3(Φ_µ,*(K_1,K_2)) V_3(K_1) V_3(K_2); a value greater than the same product for Euclidean balls would disprove Theorem 1.2. Alternatively, for a star body L and a zonal τ, check whether V_n(Ψ_τ,i L)/V_n(L)^i exceeds V_n(I_i L)/V_n(L)^i, which would contradict Corollary 5.7 or Theorem 1.5.","supporting_citations":[{"cited_title":"Affine vs. Euclidean isoperimetric inequalities","cited_arxiv_id":"1804.11165","evidence_quote":"Supplies Theorem 1.1, the black-box volume product maximization, and Lemma 3.1, the averaging identity for Φ_µ over rotations."},{"cited_title":"Schuster, Volume inequalities and additive maps of convex bodies , Mathematika 53 (2006), 211–234","cited_arxiv_id":null,"evidence_quote":"Supplies the polarization formula (1.2) establishing existence of the mixed Minkowski valuations Φ_µ and the definition of generalized centroid bodies Γ_µ."},{"cited_title":"Lutwak, Mixed projection inequalities , Trans","cited_arxiv_id":null,"evidence_quote":"Lutwak's mixed projection inequalities, the discrete case that Theorem 1.2 generalizes."},{"cited_title":"Leng and F","cited_arxiv_id":null,"evidence_quote":"Leng–Lu intersection inequalities of all orders, the results that Theorem 1.5 generalizes."},{"cited_title":"Grinberg, Isoperimetric inequalities and identities for k-dimensio nal cross-sections of convex bodies, Math","cited_arxiv_id":null,"evidence_quote":"Grinberg's affine isoperimetric inequalities for dual affine quermassintegrals, used in the right-hand bound of (1.8)."},{"cited_title":"Busemann, Volume in terms of concurrent cross-sections , Pacific J","cited_arxiv_id":null,"evidence_quote":"Busemann's intersection inequality, the classical result recovered in the dual setting."},{"cited_title":"Lutwak, Inequalities for Hadwiger’s harmonic quermassintegrals, Math","cited_arxiv_id":null,"evidence_quote":"Introduces affine quermassintegrals and Conjecture 2.1, connecting Theorem 1.4 to a major open problem."}],"review_version":1}