{"id":"c76a7259-f300-42c5-8928-ed9d428b3961","arxiv_id":"2608.05299","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For n at least 10, the open modular curve X_1(n) is isomorphic, as a scheme over Z[1/n], to the realization space of the elliptic matroid T_n.","lead":"This paper shows that a modular curve, which classifies elliptic curves with a chosen torsion point, is the same geometric object as the space of point arrangements satisfying a simple collinearity rule. It proves the identification over every field of characteristic not dividing n, and upgrades it to an isomorphism of schemes.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified.","rationale":"The reader's weakest assumption concerns background results in modular curves (Deligne–Rapoport, Katz–Mazur, Conrad, and the |3O| embedding fact), which are standard and used as black boxes. The reader's conditional verdict is driven by the unproved small-n computations in Section 10.3. Those computations are not needed for the main theorem, which is restricted to n≥10. I checked the main argument for internal inconsistencies and found none. The field-valued proof is elementary and each step can be followed: the seed pencil construction, the odd/even propagation, the irreducibility and smoothness proofs, and the group-law recovery are all sound. The scheme-theoretic proof is more terse but the key steps are present: Proposition 14.1 is a valid Artinian-point criterion; Lemma 15.3's relative Chasles argument uses Nakayama's lemma correctly; Lemma 15.6 reduces exactness of order to the special fiber, which is legitimate since n is invertible. Therefore I do not believe the central claim is at risk. The agreement_with_reader is partial because the reader's stated weakest assumption differs from the actual source of the conditional verdict, though both point to places where the paper could be more explicit.","tokens_in":32676,"tokens_out":42316,"duration_ms":395115,"concrete_test":"Run a Gröbner-basis computation in Macaulay2 or Singular to independently verify the direct computation in Section 10.3: for n=7, 8, 9, compute the Plücker ideal of T_n modulo a prime p not dividing n and check that the realization space R_{T_n} is P^1 minus 3, 4, or 5 points respectively. If this fails, only the peripheral small-n claim is affected; the n≥10 theorems would still stand.","verdict_should_be":"UNCHANGED","load_bearing_attack":"After checking the central claim (Theorem 1.2 for n≥10), I found no load-bearing flaw in the argument. The field-valued reconstruction is internally consistent: the zero-sum grid propagation uses only distinct index triples, the irreducibility and smoothness arguments are valid over arbitrary fields of characteristic prime to n, and the group-law recovery is a correct system of linear equations in the generalized Jacobian. The scheme-theoretic chain—relative seed pencil, relative Chasles via Nakayama, relative group-law recovery via the abstract Lemma 10.5, and the Artinian-point criterion—is logically sound, and the Artinian-point criterion itself is correctly applied to the finite-type morphism β. The only asserted-but-unproved computation appears in Section 10.3 for n≤9, and the paper explicitly says this range is not needed for the main theorem. Thus the reader's conditional verdict is a reasonable minor-completeness note, not a threat to the central claim.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper defines, for n≥4, the rank-3 elliptic matroid T_n on Z/nZ whose non-bases are the three-element subsets of distinct elements summing to zero. It then proves two main results. Theorem 1.1 establishes a natural bijection β_k: X_1(n)^∘(k) → R_{T_n}(k) for every n≥10 and every field k with char(k)∤n, where X_1(n)^∘ is the open part of the modular curve whose generalized elliptic curves have irreducible geometric fibers and R_{T_n}(k) is the set of rescaling classes of k-realizations of T_n. Theorem 1.2 upgrades this field-valued bijection to an isomorphism of schemes β: X_1(n)^∘ → R_n over Z[1/n]. The proof of Theorem 1.1 reconstructs, from an arbitrary realization, a unique irreducible plane cubic through all marked points using a seed pencil, Chasles propagation through zero-sum grids, and a group-law recovery argument; the cases n=10 and n=11 are handled separately. Part 2 extends the reconstruction over local Artinian rings and applies an Artinian-point criterion for isomorphisms. An appendix builds a canonical band-scheme realization space over F_1^±. As a corollary, the non-representability of T_p over Q for primes p≥11 is shown to be equivalent to the prime-order case of Mazur's theorem.","tokens_in":32841,"tokens_out":58119,"duration_ms":397647,"significance":"If correct, this is a substantial generalization of the Borisov–Roulleau theorem: it replaces complex-analytic methods by an elementary algebraic and incidence-theoretic argument, extends the correspondence to all fields of characteristic prime to n, and gives a natural Z[1/n]-model of X_1(n)^∘ as a matroid realization space. The proof is unusually self-contained in Part 1, with the n=10 and n=11 cases worked out in detail, and the scheme-theoretic Part 2 is logically coherent. I specifically checked the points that a skeptic might worry about. The Artinian-point criterion is stated and proved in Section 14.1, the relative seed-pencil and Chasles arguments in Section 15 are supplied in detail rather than assumed, and the application of the criterion in Corollary 15.8 is legitimate because every local Artinian Z[1/n]-algebra has residue field of characteristic not dividing n. The skeptical concern about a gap in the deformation-theoretic step therefore does not land. The main theorems are precisely stated and the central reconstruction argument is internally consistent.","major_comments":[],"minor_comments":[{"comment":"The assertions about the range 4≤n≤9, in particular the statement that β is an isomorphism for every field of characteristic not dividing n when 7≤n≤9, are presented with only a sketch and with reference to [8]. These claims are outside the main theorem, so they do not affect Theorems 1.1 and 1.2, but the wording 'It follows easily that β is in fact an isomorphism for every field' should be clearly marked as a summary of Borisov–Roulleau's computations or supplied with a full proof, so that the reader can distinguish new results from survey claims.","section":"Section 10.3"},{"comment":"The displayed exact sequence for restriction to a plane cubic has the wrong twist: for a cubic divisor C⊂P^2_A the kernel is O_{P^2_A}(−3), not O_{P^2_A}(−2). The conclusion is unaffected because both H^0 and H^1 vanish for either twist, so this is a typographical error rather than a gap, but it should be corrected.","section":"Lemma 15.5"},{"comment":"The reference '(Theorem 4.3)' in the second paragraph of Section 10.3 appears to be a misprint; the unique cubic through the ten-point window is proved in Corollary 4.3, and the numbering should be corrected.","section":"Section 10.3"},{"comment":"There is a missing space and period in the sentence 'entries of this grid are pairwise distinct in Z/nZEight of the nine entries lie in the window'; it should read 'in Z/nZ. Eight of the nine entries...'.","section":"Lemma 6.1"},{"comment":"The display showing the sum over i of [P^i] is malformed ('n−1X i=0'); please correct the summation notation for readability.","section":"Remark 13.1"},{"comment":"The sentence 'We do not work with Z[1/n]-schemes, rather than Z-schemes' is grammatically incomplete; it should read 'We work with Z[1/n]-schemes rather than Z-schemes...'.","section":"Introduction, Section 1.1"},{"comment":"The notation R_M is used both for the ordinary realization scheme and for the band scheme in Appendix A. Please disambiguate these objects (for example, by writing R_M^{band} for the band scheme) to avoid confusion in statements such as Proposition A.10 and Definition A.9.","section":"Appendix A and Section 11"},{"comment":"The coordinate-recovery argument at the end of the proof is quite terse, especially the sentence explaining that the Laurent monomial in the pinned units 'corrects' a displayed ratio to multidegree zero. Spelling out this correction explicitly for at least one coordinate, such as Y_i when B_i={i,1,2}, would improve readability without adding much length.","section":"Proposition 12.6"}],"recommendation":"minor_revision","confidential_remarks":"I found no load-bearing error in the central argument for n≥10. The paper is a strong fit for the journal and will be a useful reference if the minor presentation issues, especially the exact-sequence twist in Lemma 15.5 and the unproved n≤9 assertions in Section 10.3, are cleaned up. The AI-usage disclosure is unusual but transparent and does not by itself affect my assessment."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Baker proves that for n≥10 and any field k of characteristic prime to n, the open modular curve X_1(n)^∘ and the realization space of the elliptic matroid T_n are naturally bijective, and upgrades this to an isomorphism of Z[1/n]-schemes. The Borisov–Roulleau theorem over C is recovered and extended; the scheme-theoretic statement is new. The proof method is elementary—seed pencil, zero-sum grid propagation via Chasles, group-law recovery—and appears sound. I read through the n≥10 field-valued argument and found no gap; the n=10 and n=11 cases are handled separately and check out. The deformation-theoretic part (Artinian-point criterion, relative Chasles, Picard-form chord law) is a legitimate way to pass from field points to scheme isomorphism, and the reliance on Deligne–Rapoport/Katz–Mazur/Conrad is standard background rather than a hidden assumption.\n\nThe real caveat is Section 10.3: for 7≤n≤9 the paper claims realization spaces are P^1 minus 3, 4, 5 points by 'direct computation' and doesn't show the computation. That's outside the main theorem, so it's a minor completeness issue, but a referee should ask for the computation or a precise citation. Also, the paper is transparent that everything is over Z[1/n]; the integral situation is left open, which is fine.\n\nThe appendix on band schemes is a plus: it gives a canonical model over F_1^± and identifies T-points with the reduced Dressian. The connection to Mazur's theorem is correctly presented as a corollary of the correspondence, not a new proof.\n\nMy overall read: the central claim is solid, the exposition is honest, and the paper deserves a serious referee. The small-n claims should be cleaned up, and the authors should double-check the citations for the 'three-term Plücker' partial-field perfectness in the appendix, but none of this threatens the main result. If I were editor, I'd send it to review.","headline":"A solid and genuinely new scheme-theoretic extension of the Borisov–Roulleau correspondence; the central argument holds up, with only a minor completeness gap in the small-n section.","tokens_in":33343,"tokens_out":1913,"would_cite":true,"duration_ms":17826,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["05B35","14H52","14G35","14N05"],"pacs":[],"model":"deepseek-v4-flash","headline":"For every n ≥ 10, the open modular curve X_1(n)^∘ and the realization space of the elliptic matroid T_n are the same scheme over Z[1/n], making the matroid's realizations exactly the torsion configurations on smooth or nodal plane cubics.","keywords":["elliptic matroids","modular curves","matroid realization spaces","zero-sum grid propagation","plane cubic group law","Artinian deformations","scheme-theoretic isomorphism","rational torsion points"],"falsifier":"Enumerate both sides of $\\beta_k$ for a small explicit case, say $n=10$ and a finite field $k$ of characteristic not dividing $10$, and compare the number of $k$-points: a mismatch of cardinalities would disprove the field-valued bijection. Alternatively, take a normalized realization of $T_{10}$ over the dual numbers $k[\\varepsilon]/(\\varepsilon^2)$ and check whether its first-order deformation is realized by a deformation of the corresponding marked cubic; an Artinian deformation not coming from the modular side would falsify the scheme-theoretic isomorphism.","tokens_in":32447,"feed_emoji":"📐","tokens_out":15187,"duration_ms":115425,"temperature":0.7,"pith_summary":"This paper proves that two objects studied in different branches of mathematics coincide for every $n \\geq 10$: the open modular curve $X_1(n)^\\circ$, whose points parameterize elliptic curves (or irreducible nodal generalized elliptic curves) equipped with a point of exact order $n$, and the realization space $R_n$ of the elliptic matroid $T_n$, the rank-3 matroid on $\\mathbb{Z}/n\\mathbb{Z}$ whose non-bases are the three-element subsets with sum zero. For every field $k$ with $\\mathrm{char}(k) \\nmid n$, a $k$-point of the modular curve maps bijectively to a rescaling class of $k$-realizations of $T_n$, and this correspondence is upgraded to an isomorphism of schemes over $\\mathbb{Z}[1/n]$. The proof is purely algebraic and incidence-theoretic, using only projective geometry, intersection counting, and the group law on plane cubics, replacing an earlier complex-analytic comparison. A corollary reformulates the classical statement that no elliptic curve over $\\mathbb{Q}$ has a rational point of prime order $p \\geq 11$ as the non-representability of the elliptic matroid $T_p$ over $\\mathbb{Q}$.","feed_headline":"For n≥10, a modular curve is a matroid realization space","feed_subtitle":"A purely projective-geometry proof identifies X_1(n)^∘ with the realization space of the elliptic matroid over Z[1/n].","key_machinery":"The load-bearing construction is the zero-sum grid: a $3\\times 3$ array of labels in $\\mathbb{Z}/n\\mathbb{Z}$ whose rows and columns each sum to zero and whose nine entries are pairwise distinct. The row lines and column lines through the corresponding marked points form two reducible cubics; their scheme-theoretic intersection is exactly the nine marked points, and the classical theorem on cubics through eight of nine intersection points forces any cubic through eight of them to pass through the ninth. A seed pencil through nine points, with a tenth point selecting a unique member, is then propagated across $\\mathbb{Z}/n\\mathbb{Z}$ by successively applying this completion to carefully chosen grids. The group law is recovered by encoding the collinearity relations as equations in the generalized Jacobian: the line-section class $\\lambda = [H - 3P_0]$ is forced to vanish and each marked point becomes $iP_1$, with $P_1$ of exact order $n$. Scheme-theoretically, the same grid argument is run over local Artinian rings with relative lines and sections, using flatness and local algebra to promote field-level transversality, and an Artinian-point criterion upgrades the resulting bijection on all Artinian $A$-valued points to the global isomorphism.","core_discovery":"At the center of the paper is a two-way dictionary between torsion configurations and matroid realizations. Starting from a marked plane cubic $(E,O,P)$ with $P$ of exact order $n$, the points $iP$ for $i \\in \\mathbb{Z}/n\\mathbb{Z}$ form a configuration in which three distinct points are collinear exactly when their labels sum to zero in $\\mathbb{Z}/n\\mathbb{Z}$; this is a realization of $T_n$. The paper proves the reverse direction: any realization of $T_n$ over a field $k$ of characteristic not dividing $n$ determines a unique cubic through all its points, that cubic is irreducible and the marked points are smooth on it, the smooth locus carries a group law with $P_0$ as identity, and the configuration is recovered as $P_i = iP_1$ with $\\mathcal{O}_C(1) \\cong \\mathcal{O}_C(3P_0)$. The reconstruction works uniformly on local Artinian rings, which by an Artinian-point criterion yields the scheme-theoretic isomorphism $\\beta: X_1(n)^\\circ \\to R_n$ over $\\mathbb{Z}[1/n]$.","pith_inferences":["A natural testable extension would be to relax the matroid or change the modular model at primes $p$ dividing $n$: the paper shows the simple matroid realization must break there, since level sections need not give distinct points, but the isomorphism might survive in a modified form.","The reconstruction of a group law from purely collinearity data suggests a general rigidity principle: sufficiently large matroid configurations with prescribed collinearities can force an ambient algebraic-group structure, with possible analogues for other linear systems or higher-rank moduli problems.","Because the proof is purely incidence-theoretic, analogous zero-sum hypergraphs on other abelian groups could yield new modular-curve identifications or new realization-space models for related modular curves."],"forward_implications":["For every field $k$ with $\\mathrm{char}(k) \\nmid n$ and every $n \\geq 10$, the collinearity pattern of $T_n$ characterizes exactly the torsion-point configurations of smooth or nodal plane cubics of level $n$.","The modular curve $X_1(n)^\\circ$ acquires a natural affine model over $\\mathbb{Z}[1/n]$ as the multidegree-zero coordinate ring of the matroid realization scheme, so arithmetic questions about this modular curve can be translated into matroid data.","For primes $p \\geq 11$, the statement that the elliptic matroid $T_p$ is not representable over $\\mathbb{Q}$ is equivalent to the classical theorem that no elliptic curve over $\\mathbb{Q}$ has a rational point of order $p$.","The known complex-analytic comparison over $\\mathbb{C}$ follows as a special case, without modular forms or computer-assisted computation.","The realization scheme has a canonical band-scheme model over $\\mathbb{F}_1^\\pm$ whose tropical points are the reduced Dressian of $T_n$, i.e., the valuated matroids with underlying matroid $T_n$ up to rescaling."],"supporting_citations":[{"why":"the complex-analytic comparison that this paper recovers and extends.","marker":"[8]"},{"why":"supplies representability of the Γ1(n) moduli problem by X_1(n) over Z[1/n] and the generalized elliptic curve theory.","marker":"[10]"},{"why":"the arithmetic moduli theory of elliptic curves and level structures used for the integral model.","marker":"[14]"},{"why":"the formal theory of generalized elliptic curves and level structures needed for the scheme-level statements.","marker":"[9]"},{"why":"the classical theorem on cubics through eight of nine intersection points used in the propagation step.","marker":"[11]"},{"why":"the torsion theorem whose prime-order case is reformulated as non-representability of T_p over Q.","marker":"[15]"},{"why":"the flatness and Artinian-point-criterion facts used in the relative reconstruction.","marker":"[16]"}],"fun_headline_variants":["Modular curves as matroid realization spaces","Torsion points ↔ matroid realizations: a new dictionary","X_1(n) is a matroid realization space over Z[1/n]","Elliptic matroids capture modular curves"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument presupposes the standard theory that the $\\Gamma_1(n)$ moduli problem is representable by the smooth scheme $X_1(n)$ over $\\mathbb{Z}[1/n]$ and that on its open locus $X_1(n)^\\circ$ the universal generalized elliptic curve is embedded in the projective plane by the complete linear system $|3O|$; if either the representability or the embedding failed, the morphism $\\beta$ and the reconstruction would have no starting point.","fun_headline_variants_meta":{"raw":{"variants":["Modular curves as matroid realization spaces","Torsion points ↔ matroid realizations: a new dictionary","X_1(n) is a matroid realization space over Z[1/n]","Elliptic matroids capture modular curves"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000175,"raw_usage":{"total_tokens":1370,"prompt_tokens":1114,"completion_tokens":256,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":730,"completion_tokens_details":{"reasoning_tokens":188}},"tokens_in":730,"tokens_out":256,"duration_ms":3373,"temperature":1.0,"reasoning_tokens":188,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T15:52:52.648917+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Enumerate both sides of $\\beta_k$ for a small explicit case, say $n=10$ and a finite field $k$ of characteristic not dividing $10$, and compare the number of $k$-points: a mismatch of cardinalities would disprove the field-valued bijection. Alternatively, take a normalized realization of $T_{10}$ over the dual numbers $k[\\varepsilon]/(\\varepsilon^2)$ and check whether its first-order deformation is realized by a deformation of the corresponding marked cubic; an Artinian deformation not coming from the modular side would falsify the scheme-theoretic isomorphism.","supporting_citations":[{"cited_title":"Deligne and M","cited_arxiv_id":null,"evidence_quote":"supplies representability of the Γ1(n) moduli problem by X_1(n) over Z[1/n] and the generalized elliptic curve theory."},{"cited_title":"Katz and B","cited_arxiv_id":null,"evidence_quote":"the arithmetic moduli theory of elliptic curves and level structures used for the integral model."},{"cited_title":"Conrad,Arithmetic moduli of generalized elliptic curves, J","cited_arxiv_id":null,"evidence_quote":"the formal theory of generalized elliptic curves and level structures needed for the scheme-level statements."},{"cited_title":"Eisenbud, M","cited_arxiv_id":null,"evidence_quote":"the classical theorem on cubics through eight of nine intersection points used in the propagation step."},{"cited_title":"Mazur,Modular curves and the Eisenstein ideal, Inst","cited_arxiv_id":null,"evidence_quote":"the torsion theorem whose prime-order case is reformulated as non-representability of T_p over Q."}],"review_version":1}