{"id":"fb0d58a6-f77f-4ee4-a602-c28fa783c9c0","arxiv_id":"2506.22415","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Covolume polynomials are characterized as the differential operators that preserve volume polynomials, and the resulting product theorem implies that intersections of algebraic matroids are algebraic.","lead":"The paper proves that Aluffi's covolume polynomials are exactly the polynomial differential operators that preserve volume polynomials, closing a known gap in earlier work. This characterization yields new results in matroid theory, including that the intersection of two algebraic matroids is again algebraic over the same field.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.3 rests on the unrefereed [HHM+, Lemma 2.6] imported in Lemma 2.1; if that lemma fails or has hidden hypotheses, the central characterization and its matroid applications collapse, so the paper is correctly conditional.","rationale":"I read the paper in good faith and followed the main chain: Theorem 1.3 reduces preserving realizable volume polynomials to the intersection geometry of gY and Z; the coefficient-to-intersection computation itself is transparent and I found no internal inconsistency there. The proof of Theorem 1.5 from Theorem 1.3 is formal and sound. The converse direction of Theorem 1.3 is effectively by definition, and the symbol theorem and its corollaries are used consistently. The one point where the whole central claim becomes hostage to material outside this text is Lemma 2.1, which imports [HHM+, Lemma 2.6] verbatim and uses it again in Theorem 2.7 and Proposition 2.10. That is also the weakest point identified by the Reader, and I agree with that assessment. I do not see a separate internal flaw that would change the verdict. The paper's self-disclosure about the [RS]/[GS] gap is a sign of good scholarly practice, and the use of two external results from the same concurrent preprint is explicit rather than hidden. Still, because the main characterization is conditional on a not-yet-refereed lemma, the Reader's CONDITIONAL verdict is appropriate; my stress-test pass does not move it.","tokens_in":21561,"tokens_out":34980,"duration_ms":430434,"concrete_test":"Verify [HHM+, Lemma 2.6] in the exact setting of Lemma 2.1, or prove Lemma 2.1 directly: restrict p: VY,Z → Ḡ to a general curve C in Ḡ, form the Stein factorization V_C → C′ → C, and check that the components of a general fiber are fibers of V_C → C′ over two points of the curve C′. Since all points of a complete curve are algebraically equivalent, this would establish the needed algebraic equivalence and close the gap. If this direct proof succeeds, the objection is resolved; if it reveals a missing hypothesis such as flatness or a complete base, the main theorem is at risk.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central characterization (Theorem 1.3) is proved through Lemma 2.1, whose proof quotes [HHM+, Lemma 2.6] as a black box: for a proper surjective morphism between irreducible varieties, the irreducible components of a general fiber are algebraically equivalent in the source. The same lemma is reused in Theorem 2.7 and Proposition 2.10. This input comes from the authors' concurrent, not yet refereed preprint arXiv:2505.08881 and is not proved in this paper. If Lemma 2.6 is false, or if it requires hypotheses not verified here (for example flatness of the morphism or completeness of the base), then the step in Theorem 1.3 where a component V of gY ∩ Z is assigned a cycle class proportional to Σ e_α [P_α] in CH(P_μ)⊗Q is unsupported; consequently Theorem 1.3, Corollary 1.4, Theorem 1.5, and the matroid intersection theorem 5.11 would not follow. A secondary unstated premise is that algebraic equivalence of cycles, after pushforward to a product of projective spaces P_μ, implies equality of rational equivalence classes. This is true for P_μ because the cycle class map to cohomology is an isomorphism, but it is not said. The external lemma is the genuinely load-bearing weakness: it is the one place where correctness of this paper's main theorem is outsourced to an unreviewed concurrent preprint.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves that Aluffi's covolume polynomials are exactly the polynomial differential operators that preserve realizable volume polynomials over an algebraically closed field k (Theorem 1.3), with a limit version for volume polynomials (Corollary 1.4). It also proves a symbol theorem for realizable volume polynomials (Theorem 1.12), a dual characterization of volume polynomials via operators preserving covolume polynomials (Theorem 1.9), and several operator-preservation statements. The final section applies these results to algebraic polymatroids, culminating in Theorem 5.11: the intersection of two matroids algebraic over k is algebraic over k. The central proof realizes volume polynomials as classes of subvarieties of products of projective spaces, applies Kleiman transversality to a general translate, and uses a lemma on algebraic equivalence of the components of the intersection to identify the result of the differential operator with the volume polynomial of a single component.","tokens_in":21739,"tokens_out":12787,"duration_ms":137548,"significance":"If the main theorem is correct, it gives a clean homological characterization of covolume polynomials and strengthens Aluffi's results on closure under multiplication and nonnegative linear changes of coordinates. The symbol theorem and the operator-preservation results are natural and useful analogues of the Lorentzian-polynomial theory, and the matroid intersection theorem is a substantial new application. The term-by-term computation in the proof of Theorem 1.3 is careful and the geometric framework is coherent. The main weakness is that the proof of Lemma 2.1, which is load-bearing for the entire characterization, is outsourced to a lemma from the concurrent unreviewed preprint [HHM+], so the central theorem is conditional on an external result whose proof is not included.","major_comments":[{"comment":"The proof of Lemma 2.1 applies [HHM+, Lemma 2.6] as a black box: the assertion that the irreducible components of a general fiber of a proper surjective morphism between irreducible varieties are algebraically equivalent is not proved or even stated in this manuscript. This lemma is used in the proof of Theorem 1.3 and again in Theorem 2.7 and Proposition 2.10, so if [HHM+, Lemma 2.6] has hidden hypotheses or is false, the characterization and the matroid applications collapse. The manuscript should either include a full proof of the needed statement or cite a refereed version; relying on an unreviewed concurrent preprint is not acceptable for a load-bearing step.","section":"Section 2, Lemma 2.1"},{"comment":"After Lemma 2.1, the proof concludes that any irreducible component V of gY∩Z satisfies λ1λ2λ3[V] = Σ e_α[P_α]. This inference uses the fact that algebraic equivalence of cycles on P_μ implies equality of rational equivalence classes; this is true over Q because the cycle class map on P_μ is an isomorphism, but the step is not stated. Since this is the bridge from the geometric intersection to the polynomial g(B)∘f, it should be made explicit.","section":"Section 2, proof of Theorem 1.3"},{"comment":"The proof of Proposition 4.4 uses [HHM+, Theorem 1.8] to assert that every quadratic Lorentzian polynomial with rational coefficients is a realizable volume polynomial. This is another import from the same unreviewed preprint. The proposition is an advertised application, so the dependency should be addressed: either prove the quadratic characterization or cite a published reference for it.","section":"Section 4, Proposition 4.4"}],"minor_comments":[{"comment":"The sentence 'Since πY,Z is open and surjective, π−1(y×z) is irreducible for every closed point y×z' is misleading; irreducibility of the fibers follows from flatness together with the irreducibility of the model fibers, not from openness and surjectivity. Please rephrase.","section":"Section 2, proof of Lemma 2.1"},{"comment":"In the purely transcendental case, the sentence 'Since the polynomial ring R[x] is an integral domain for any integral domain R, it follows that Xℓ is reduced and irreducible' does not by itself explain why the base change of the subvariety is irreducible; a direct argument using that k[t] is a domain, or a reference for base change by a purely transcendental extension, would be clearer.","section":"Section 2, Proposition 2.10"},{"comment":"There is a typo: 'specturm' should be 'spectrum'.","section":"Section 5, Proposition 5.4"},{"comment":"The remark that any coefficient inequality for volume polynomials can be applied to s_w(B)∘f(x) is clear in spirit, but it would be helpful to indicate that the new polynomial is itself a volume polynomial and not merely a polynomial with nonnegative coefficients.","section":"Section 1, Remark 1.8"}],"recommendation":"major_revision","confidential_remarks":"The central theorem is attractive and the overall strategy is convincing, but the dependence of Lemma 2.1 on the concurrent unreviewed preprint [HHM+] is substantial. I would not accept the paper in its current form until that dependency is resolved, either by including a complete proof of the needed algebraic-equivalence lemma or by the companion paper passing review. The remaining issues are local and can be fixed by clarification."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The short version: this is a serious paper with a real main theorem, and the only serious weakness is that its central geometric engine is outsourced to an unrefereed companion preprint. Conditional is the right verdict, if by conditional you mean \"send to referees and make them check the companion.\"\n\nWhat is new and what is done well: Theorem 1.3 is exactly the kind of statement that should be true and wasn't known: Aluffi's covolume polynomials are precisely the differential operators preserving realizable volume polynomials. The proof structure is transparent. I verified the key coefficient-to-intersection computation in Theorem 1.3 and the support computation for Theorem 5.11; both are sound. The paper also closes a gap in [RS]/[GS] that the authors themselves flag in Remark 1.7, which is honest. The matroid intersection application (Theorem 5.11) is genuinely new and follows cleanly from the product closure of covolume polynomials.\n\nWhere the soft spots are: Lemma 2.1 is load-bearing. It imports [HHM+, Lemma 2.6] from a concurrent preprint by the same group, and that lemma is used in Theorem 1.3, Theorem 2.7, and Proposition 2.10. If that lemma fails, the main characterization and the matroid theorem do not follow. This is a real vulnerability, not a cosmetic one. The secondary point about algebraic equivalence implying equality of cycle classes in CH(P_mu) is true for P_mu but not stated; that's a minor exposition fix. Also, the converse direction of Theorem 1.3 is close to definitional—that's not a flaw, just an observation that the content is entirely in the forward direction.\n\nBottom line: if the companion preprint is correct, this paper is a substantial contribution to the volume/covolume theory and to algebraic matroids. The dependence on an unrefereed companion is uncomfortable but normal in this area; the authors are the right people to have made the connection. Send it to a serious referee. The referee should have access to arXiv:2505.08881 and should be asked to verify Lemma 2.6 and its hypotheses. I would bring it to a reading group.","headline":"Genuinely new and mostly sound, but the main theorem's geometric engine is imported from an unrefereed companion preprint; conditional is the right call and it deserves referees.","tokens_in":22401,"tokens_out":2060,"would_cite":true,"duration_ms":20842,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14C25","05B35","14M15"],"pacs":[],"model":"deepseek-v4-flash","headline":"A polynomial differential operator preserves realizable volume polynomials exactly when it is a realizable covolume polynomial, and the dual analysis shows volume polynomials are the operators that preserve covolume polynomials.","keywords":["volume polynomials","covolume polynomials","differential operators","Lorentzian polynomials","algebraic matroids","polymatroids","M-convex sets","symbol theorem"],"falsifier":"Look for a proper surjective morphism from an irreducible variety to a curve whose general fiber has two irreducible components whose cycle classes in the Chow group of the total space are not algebraically equivalent; such an example would invalidate the geometric lemma and with it the engine of Theorem 1.3.","tokens_in":21237,"feed_emoji":"🧮","tokens_out":9471,"duration_ms":92202,"temperature":0.7,"pith_summary":"This paper classifies the polynomial differential operators that preserve volume polynomials. A volume polynomial records the normalized growth of Minkowski sums of convex bodies, or of top self-intersections of semiample divisors on a projective variety; a covolume polynomial is a polynomial in differential operators whose application to the test monomial $x^{[\\mu]}$ is a volume polynomial. The main theorem states that a polynomial operator preserves every realizable volume polynomial over $k$ if and only if it is itself a realizable covolume polynomial over $k$, with the same statement for limits. The dual theorem classifies volume polynomials as exactly the operators that send covolume polynomials to covolume polynomials. Concrete payoffs include a symbol theorem for such operators, a characterization of algebraic polymatroids as supports of realizable volume polynomials, and the result that the intersection of two matroids algebraic over $k$ is algebraic over $k$.","feed_headline":"Covolume polynomials = operators that preserve volume polynomials","feed_subtitle":"The duality also makes the intersection of two algebraic matroids algebraic, not just the union.","key_machinery":"The machinery has two parts. First, the basic construction encodes a realizable volume polynomial $f=\\sum_\\alpha c_\\alpha x^{[\\alpha]}$ as the Chow class of a subvariety of a product of projective spaces, $\\sum_\\alpha c_\\alpha [\\mathbb{P}_\\alpha]$, so that statements about polynomials become statements about algebraic cycles. Second, the geometric engine is Lemma 2.1: for proper morphisms from irreducible varieties to a complete homogeneous variety, the irreducible components of a general fiber product are algebraically equivalent in the total space. This lets the paper pass from a general translate intersection, whose components may be many and disconnected, to a single component carrying the same cycle class, which is what makes the differential-operator statement follow from intersection theory.","core_discovery":"Theorem 1.3 asserts that for any $g\\in \\mathbb{Q}[\\mathbf{B}]$, the following are equivalent: $g$ is a realizable covolume polynomial over $k$, and $g(B)\\circ f(x)$ is a realizable volume polynomial over $k$ for every realizable volume polynomial $f$ over $k$. The paper proves this by showing the general-preservation direction geometrically and the converse by testing on the single realizable volume polynomial $x^{[\\mu]}$; since the monomial test is part of the definition of covolume, the two conditions coincide. Taking limits gives the same equivalence between covolume polynomials and operators preserving all volume polynomials, and Theorem 1.9 proves the dual statement: a polynomial $f$ is a realizable volume polynomial exactly when multiplication by $f$ sends every realizable covolume polynomial to a realizable covolume polynomial. From these statements the paper derives a symbol theorem for linear operators on polynomial rings and a product rule for realizable covolume polynomials, which in turn yields the matroid intersection theorem.","pith_inferences":["Beyond the paper: the one-monomial characterization gives a certification rule: to prove an operator preserves all realizable volume polynomials, one only has to check its symbol against the single monomial $x^{[\\mu]}$.","Beyond the paper: if the same operator-product logic could be run with the volume-polynomial product rule alone, the dual of an algebraic matroid would be algebraic over $k$; the paper's Theorem 5.11 stops short of this long-open question.","Beyond the paper: the analytic analogue using semipositive classes on compact Kähler manifolds, which the final remark proposes, could be tested by checking whether the algebraic-equivalence lemma holds in that setting; a failure there would separate algebraic from analytic matroids over $\\mathbb{C}$."],"forward_implications":["The product of realizable covolume polynomials over $k$ is a realizable covolume polynomial, and the product of realizable volume polynomials over $k$ is a realizable volume polynomial.","Every nonnegative rational linear change of variables takes realizable covolume polynomials to realizable covolume polynomials.","If the symbol of a homogeneous linear operator is a realizable volume polynomial, then the operator sends realizable volume polynomials to realizable volume polynomials; this yields preservation statements for polarization, normalization, interlacing, and symmetric exclusion operators.","A polymatroid is algebraic over $k$ if and only if it is the support of a realizable volume polynomial over $k$.","The intersection of two matroids algebraic over $k$ is algebraic over $k$; the same holds for minors, truncations, and Higgs lifts."],"supporting_citations":[{"why":"Supplies the lemma that irreducible components of a general fiber of a proper surjective morphism are algebraically equivalent, used in Lemma 2.1 and Proposition 2.10, and the quadratic Lorentzian characterization used in Proposition 4.4.","marker":"[HHM+]"},{"why":"Introduces covolume polynomials and earlier closure results that the paper strengthens to the realizable and limit settings.","marker":"[Alu24]"},{"why":"Provides the Lorentzian polynomial framework, the M-convex support theorem for volume polynomials, and the Lorentzian symbol theorem that the volume symbol theorem parallels.","marker":"[BH20]"},{"why":"Identifies bivariate Lorentzian forms with realizable volume polynomials, used in the truncation, normalization, and interlacing arguments.","marker":"[Huh12]"},{"why":"Relates algebraic polymatroids to transcendence degrees and gives the support statement used in Proposition 5.4.","marker":"[CCRL+20]"},{"why":"Supplies the transversality theorem ensuring a general translate intersection is equidimensional of expected dimension in the proof of Theorem 1.3.","marker":"[Kle74]"},{"why":"Gives the dually Lorentzian theory and the symbol formula adapted in the proof of Theorem 1.9.","marker":"[RSW]"},{"why":"Provides the truncation theorem used in Theorem 5.11 to reduce the non-expected-rank case of matroid intersection.","marker":"[Cun79]"}],"fun_headline_variants":["Covolume polynomials are the exact operators preserving volume polynomials","Duality: covolume polynomials act as volume polynomial preservers","Volume-preserving operators equal covolume polynomials","Matroid intersection follows from covolume-volume duality"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument rests on an imported geometric lemma: for any proper surjective algebraic map, the irreducible pieces of a general fiber must be interchangeable by an algebraic deformation; if this fails, the main characterization collapses.","fun_headline_variants_meta":{"raw":{"variants":["Covolume polynomials are the exact operators preserving volume polynomials","Duality: covolume polynomials act as volume polynomial preservers","Volume-preserving operators equal covolume polynomials","Matroid intersection follows from covolume-volume duality"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000728,"raw_usage":{"total_tokens":3183,"prompt_tokens":792,"completion_tokens":2391,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":408,"completion_tokens_details":{"reasoning_tokens":2324}},"tokens_in":408,"tokens_out":2391,"duration_ms":19115,"temperature":1.0,"reasoning_tokens":2324,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T22:12:04.657287+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Look for a proper surjective morphism from an irreducible variety to a curve whose general fiber has two irreducible components whose cycle classes in the Chow group of the total space are not algebraically equivalent; such an example would invalidate the geometric lemma and with it the engine of Theorem 1.3.","supporting_citations":[],"review_version":1}