{"id":"8694da3b-86c5-4d93-ad6a-3abfaef9672f","arxiv_id":"2507.04474","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The minimal dimension of a faithful linear representation of all finite subgroups of the two-dimensional Cremona group is 8 when the field contains a primitive cube root of unity, 6 otherwise, and infinite in positive characteristic.","lead":"This paper defines the representation dimension of finite subgroups of Cremona groups and determines it exactly in ranks 1 and 2 over every field. It also shows the invariant is infinite in positive characteristic and grows at least exponentially in the rank over characteristic 0 fields.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 8.3 leaves degree-6 del Pezzo surfaces uncovered: Corollary 7.5 gives fdim≤7 there, not 6, so the upper bound c2(k)≤6 in Theorem 1.2 has a missing case.","rationale":"The reader's weakest_assumption is the unshipped Sage computation behind the lower bound, and the reader also flags a patchable gap in the characteristic-2 descent of Theorem 1.1. Neither point directly threatens the central value c2(k)=6 as much as the degree-6 del Pezzo gap. The characteristic-2 issue does not affect Theorem 1.2, which is characteristic 0, and the Sage computation is a reproducibility concern that can be settled by shipping the script. By contrast, Proposition 8.3 explicitly cites Corollary 7.5 for d ≤ 6, and Corollary 7.5 gives d+1 for d=6, i.e. 7, so the proof of the upper bound c2≤6 has a genuine missing case. This is internal to the argument, not a question of external verification. The gap is very likely patchable because the degree-6 del Pezzo is a toric surface with six Cox generators and its automorphism group should embed into GL_6, but the paper does not supply that argument. Since the theorem is probably true and the gap is repairable, I keep the reader's CONDITIONAL verdict, adding repair of the degree-6 case to the list of conditions.","tokens_in":22712,"tokens_out":20349,"duration_ms":236685,"concrete_test":"Settle the degree-6 case by taking X = Bl_{p1,p2,p3}(P^2) over k without √−3, whose automorphism group is (G_m)^2 ⋊ S_3. Construct the six-dimensional representation of Aut(X) on the vector space spanned by the Cox ring variables, namely the six exceptional curves of the toric model. Check that the kernel of this representation is trivial for every finite subgroup G; if so, insert this as the missing case in Proposition 8.3. If some finite subgroup has no faithful representation of dimension ≤ 6, then Theorem 1.2's value “6” for fields without √−3 is false and must be revised upward.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof of the upper bound of Theorem 1.2 reduces to bounding fdim_k(Aut(X)) for minimal rational G-surfaces X. Proposition 8.3 claims that if X is a rational del Pezzo surface not isomorphic to P^2, then fdim_k(Aut(X)) ≤ 6. The argument handles degrees 8 and 7 by equivariant blowdown and then says “the remaining cases are of degree d ≤ 6, which follow from Corollary 7.5.” But Corollary 7.5 states the bound is d+1 for d ≥ 3, so for d=6 it gives 7, not 6. Degree-6 del Pezzo surfaces are rational and can be k-forms over fields without √−3, exactly the regime where the theorem claims c2=6. The missing case is not covered by another proposition in §8: the conic bundle propositions apply only when Theorem 8.1 selects a G-conic bundle with Pic(X)^G = Z^2, which is not guaranteed for an arbitrary finite subgroup. Therefore, as written, the sharp upper bound c2≤6 is not established; a separate argument, such as the six-generator Cox ring of the degree-6 del Pezzo, is required.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper defines c_n(k) as the supremum over finite subgroups G of Cr_n(k) of the minimal dimension of a faithful linear representation of G over k. It determines c_1(k) for all fields, determines c_2(k) explicitly (∞ in positive characteristic, 8 if √−3∈k, 6 otherwise), proves that c_n(k)=∞ for n≥2 in positive characteristic, proves finiteness of c_n(k) for many characteristic-0 fields, and gives exponential lower bounds in n. The proofs use minimal rational G-surfaces, algebraic tori and their symmetric ranks, weighted projective spaces, and a GAP-checked group-theoretic appendix.","tokens_in":22979,"tokens_out":26134,"duration_ms":271338,"significance":"The exact evaluation of c_2(k) is a major step in understanding finite subgroups of plane Cremona groups over arbitrary fields, and the lower-bound technology via symmetric ranks of root lattices is original and useful. The positive-characteristic non-linearity result for Cr_n(k) (n≥2) and the use of model-theoretic compactness are elegant. The appendix's hand proof of the six-dimensional bounds for extensions of polyhedral groups, accompanied by GAP verification, is a valuable resource. If the gaps identified below are repaired, the paper will be a strong contribution.","major_comments":[{"comment":"Proposition 8.3 claims fdim_k(Aut(X)) ≤ 6 for every rational del Pezzo surface X not isomorphic to P^2. The proof handles degrees 8 and 7 by equivariant blowdown, then states that the remaining cases of degree d ≤ 6 follow from Corollary 7.5. However, Corollary 7.5 gives fdim_k(Aut(X)) ≤ d+1 for d ≥ 3, so for d=6 it yields the bound 7, not 6. Degree-6 del Pezzo surfaces are rational and can be k-forms over fields without √−3, precisely in the regime where Theorem 1.2 asserts c_2(k)=6; moreover Theorem 8.1 permits a minimal G-surface of this type with Pic(X)^G ≅ Z. The upper bound c_2(k) ≤ 6 is therefore not proven as written. A separate argument for the degree-6 del Pezzo case (e.g., using the structure of its anticanonical ring) is required.","section":"§8, Proposition 8.3 and proof of Theorem 1.2"},{"comment":"In the case where ρ is not absolutely irreducible and b=0 for every g∈G, the proof asserts that since ρ is defined over L∩K and finite fields are perfect, 'K = k', and hence ρ is defined over k. This inference is invalid: K/k may be a non-trivial purely inseparable extension while K contains a finite subfield such as L∩K. The equality K=k does not follow, and the descent of ρ to a representation over k is not established. Thus the upper bound c_1(k) ≤ 2 for infinite imperfect fields of characteristic 2 rests on an unsupported step. Please supply a correct descent argument or an alternative proof for this case.","section":"§3, proof of Theorem 1.1, characteristic-2 descent"},{"comment":"The values symrank(L/dL,W)=|Ω| for n=1,...,6 in Table 1 are justified by the sentence 'This was checked using [Sage]', with no code or output included. This finite verification is load-bearing: it underpins the lower bounds in Theorem 1.5 (including c_2(k) ≥ 6 in Theorem 1.2). Please provide the Sage script and its output, or give a short hand-checkable certificate (e.g., listing the orbit sizes and a generating orbit union for each lattice).","section":"§5, proof of Theorem 5.4"}],"minor_comments":[{"comment":"The term 'relative closure of the prime field F_2 in F' is not defined; please make explicit that it is the algebraic closure of F_2 inside F (a finite field).","section":"§3, proof of Theorem 1.1"},{"comment":"The notations gS4+ and gS4− for the binary octahedral groups are used without definition; please define them and explain the GAP ID column.","section":"Table 2, Appendix A"},{"comment":"The condition (b) in the proof appears garbled: the displayed expression with 'W_i' and 'W_{i≠j}' should be a disjunction expressing that the matrix (x^g_{ij}) is not the identity. Please correct the notation.","section":"§4, Lemma 4.5"},{"comment":"In the reducible case, the claim that the representation σ∨⊗τ factors through eG→G and is faithful is stated without justification; adding one sentence on why the kernel is exactly the central scalars would help.","section":"§6, Proposition 6.4"},{"comment":"The assertion that a rational k-form of P(1,1,n,n) is isomorphic to the standard weighted projective space is not immediate; please include a reference or a short justification.","section":"§8, Proposition 8.6"}],"recommendation":"major_revision","confidential_remarks":"The three major gaps are localized and appear fixable: the degree-6 del Pezzo case and the characteristic-2 descent need new arguments, and the Sage computation should be made reproducible. I recommend major revision rather than rejection because the overall framework and most of the theorems are well supported, and the missing pieces are concrete rather than conceptual."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things: the paper introduces a natural invariant c_n(k) and pins it down for n=1,2 over all fields, and one of those exact values is not fully proved as written. The stress-test note is right. Proposition 8.3 leaves degree-6 del Pezzo surfaces uncovered: Corollary 7.5 gives fdim ≤ d+1, which is 7 for d=6, not 6. Since Theorem 8.1 can land on a minimal del Pezzo surface of degree 6 with Pic(X)^G = Z, and the conic bundle propositions don't cover that case, the claimed upper bound c2(k) ≤ 6 lacks proof there. This is load-bearing, though it looks patchable: the six-generator Cox ring of a degree-6 del Pezzo should give fdim ≤ 6 with some additional argument. As written, Theorem 1.2's sharp upper bound is not established.\n\nWhat is genuinely new: the invariant itself, the exact c1(k) for all fields, the positive-characteristic infinitude via the non-linearity of k[x]⋊ k (the Mal'cev compactness argument is clever), and the B_n weight-lattice construction for the exponential lower bounds. The paper is careful about k-forms and twisting, and the group-theoretic appendix gives a full hand proof alongside GAP checks. That is solid work.\n\nOther soft spots are smaller. The characteristic 2 descent in Theorem 1.1 contains a suspicious inference (\"finite fields are perfect, so K = k\") that the reader flagged; it is likely patchable but the line is not valid as written. The lower-bound table depends on an unshipped Sage verification; the finite orbit-union check is easy enough for a referee to reproduce, but the authors should post the code.\n\nNet: this is a serious paper with a real missing case in its headline theorem. Send it to referees, but expect a revision that either fixes the degree-6 del Pezzo bound or restricts the statement of Theorem 1.2.","headline":"A genuinely valuable paper on a new invariant, with a real missing case in the main upper bound for c2(k) that a referee should catch.","tokens_in":23493,"tokens_out":2117,"would_cite":true,"duration_ms":22696,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14E07","20C99"],"pacs":[],"model":"deepseek-v4-flash","headline":"The plane Cremona group's finite subgroups need exactly 6 dimensions over most characteristic-0 fields, 8 if √−3 is present, and no finite bound in positive characteristic.","keywords":["Cremona groups","representation dimension","finite subgroups","birational automorphisms","algebraic tori","minimal rational surfaces","positive characteristic","projective linear groups"],"falsifier":"Run the orbit enumeration for the G₂ root lattice with its Weyl group acting on L/4L: if any W-stable subset of L/4L of size at most 5 generates the module, then the claimed lower bound c₂(k) ≥ 6 fails; the same check applies to A₃, F₄, D₅, and E₆ for the values 12, 24, 40, and 72.","tokens_in":22531,"feed_emoji":"🎯","tokens_out":7973,"duration_ms":75289,"temperature":0.7,"pith_summary":"This paper asks how many dimensions are needed to faithfully represent every finite subgroup of the Cremona group—the birational automorphisms of projective n-space—over a given field, and answers this exactly for n=1 and n=2 over every field. The plane Cremona group Cr₂(k) behaves differently depending on the field: over any characteristic-0 field that does not contain √−3, six dimensions suffice and are necessary; over fields containing √−3, eight dimensions are needed and suffice; and over fields of positive characteristic, no finite bound exists. For higher n, the paper proves finiteness for many characteristic-0 fields, including number fields and the complex numbers, and gives lower bounds that grow exponentially in n. The exact values come from a case analysis of rational surfaces with group actions, combined with new bounds on the representation dimensions of automorphism groups of tori and projective spaces.","feed_headline":"Plane Cremona finite groups: 6 dimensions, or 8 with √-3","feed_subtitle":"Exact value over every field; in positive characteristic, no finite bound exists.","key_machinery":"The load-bearing structure is the reduction of finite subgroups of Cr₂(k) to automorphism groups of minimal rational surfaces—del Pezzo surfaces and conic bundles—following Manin and Iskovskikh. Around that hub, the argument uses three further pieces: (1) representations of algebraic tori, where certain finite groups are realized as subgroups of Aut(T) for a rational torus T built from a root lattice, turning a combinatorial 'symmetric rank' computation into lower bounds on representation dimension; (2) the projective linear groups PGL₃(k) and PGL₄(k), where the Weil representation of (C_p)² ⋊ SL₂(F_p) gives the sharp lower bound p²−1 and a Blichfeldt classification of primitive finite subgroups gives the upper bound when p=3 and √−3 ∉ k; and (3) Mal'cev's compactness theorem, which lets the positive-characteristic result be proved by showing the group k[x]⋊k is not linear over any field.","core_discovery":"On the paper's own terms, the central discovery is the exact value of c₂(k), the supremum of minimal faithful-representation dimensions over finite subgroups of the plane Cremona group: it is 6 for any field of characteristic 0 that does not contain √−3, 8 for characteristic-0 fields that do, and infinite in positive characteristic. The lower bound of 6 is obtained by constructing, from root lattices of types A₁, G₂, A₃, F₄, D₅, E₆ and B_n, rational tori whose automorphism groups contain finite groups with large representation dimension. The upper bound of 6 or 8 uses the Manin–Iskovskikh reduction: any finite subgroup of Cr₂(k) embeds in the automorphism group of a minimal rational surface, either a del Pezzo surface or a conic bundle surface, and the paper bounds the representation dimension of each such automorphism group in turn. The remaining gap between 6 and 8 is exactly whether k contains √−3, which controls whether PGL₃(k) contains the Hessian group (C₃)² ⋊ SL₂(F₃).","pith_inferences":["The exponential lower bound 2ⁿ suggests the true growth of c_n(k) may be much faster, possibly driven by symmetric ranks of Weyl groups; whether c_n(C) is finite for all n and how it grows remains open, since the paper's finiteness proof is non-constructive.","The positive-characteristic proof shows k[x]⋊k is not linear over any field, a fact that may extend to other elementary birational groups and suggests finite subgroups of Cremona groups over finite fields still have unbounded representation dimension over the same field.","The unshipped Sage computation that certifies the lower bounds is a testable artifact: re-running the orbit enumeration for the G₂, A₃, F₄, D₅, and E₆ root lattices would independently confirm or correct the 6-, 12-, 24-, 40-, and 72-dimensional lower bounds.","The same torus-and-root-lattice method likely applies to other groups of birational transformations, such as automorphism groups of rational varieties with torus actions, where symmetric rank may again control representation dimension."],"forward_implications":["For every field k of characteristic 0, finite subgroups of the plane Cremona group are uniformly representable in at most 8 dimensions, and exactly 6 unless √−3 ∈ k.","In positive characteristic, there are finite subgroups of Cr_n(k) for every n ≥ 2 with arbitrarily large minimal faithful-representation dimension, so no finite uniform bound exists.","The three-dimensional complex Cremona group satisfies 15 ≤ c₃(C) ≤ 62208, giving the first nontrivial bounds of their kind for rank 3.","For characteristic-0 fields containing all roots of unity, c_n(k) is finite for every n, and the same holds for fields finitely generated over Q.","The lower bounds c_n(k) ≥ 2ⁿ for n ≥ 7 show that, on a fixed field, the minimal faithful-representation dimension of finite subgroups grows at least exponentially in the rank of the Cremona group."],"supporting_citations":[{"why":"Supplies the classification of finite subgroups of the plane Cremona group and the reduction to minimal rational G-surfaces (Theorem 8.1), the backbone of the c₂(k) proof.","marker":"[DI09a]"},{"why":"Provides the Jordan property for Cremona groups in characteristic 0, used for the finiteness in Theorem 1.4 and the abelian-subgroup bound in Theorem 9.2.","marker":"[PS16]"},{"why":"Mal'cev's compactness theorem powers Lemma 4.5, allowing the positive-characteristic non-linearity proof to pass from finite subgroups to a full linear representation.","marker":"[Mal40]"},{"why":"Blichfeldt's classification of primitive finite subgroups of PGL₃(C) is the key upper-bound input in Proposition 6.4 when √−3 is absent.","marker":"[Bli17]"},{"why":"Gives the list of finite subgroups of PGL₂(K) used in Theorem 1.1 and throughout the polyhedral-group appendix for conic bundle automorphism groups.","marker":"[Bea10]"},{"why":"The second author's thesis on representation dimensions of algebraic tori provides the symmetric-rank values and the conjecture underlying the lower-bound table in Theorem 5.4.","marker":"[Hea24]"},{"why":"Standard reference for root systems and Weyl groups, supplying the lattices L and Weyl groups W used in the Theorem 5.4 lower-bound constructions.","marker":"[Hum75]"},{"why":"Establishes the earlier bound c₂(C) ≤ 48 that this paper sharpens to the exact value of 6 or 8.","marker":"[Ure21]"},{"why":"The finite check that no union of orbits of size less than |Ω| generates L/4L for the relevant root lattices, which certifies the lower bounds of Theorem 5.4.","marker":"[Sage]"}],"fun_headline_variants":["Plane Cremona finite groups: 6 dims, 8 if field has √-3","Finite Cremona subgroups: exact 6 or 8 dims for plane, infinite in positive char","6 or 8 dims for plane Cremona finite groups, or infinite","Plane Cremona: finite-group rep dim is 6, 8, or infinite"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The 6-dimensional lower bound for c₂(k) rests on an unshipped Sage computation asserting that, for the root lattices G₂, A₃, F₄, D₅, and E₆, no union of orbits of size less than the claimed minimum generates L/4L (Theorem 5.4).","fun_headline_variants_meta":{"raw":{"variants":["Plane Cremona finite groups: 6 dims, 8 if field has √-3","Finite Cremona subgroups: exact 6 or 8 dims for plane, infinite in positive char","6 or 8 dims for plane Cremona finite groups, or infinite","Plane Cremona: finite-group rep dim is 6, 8, or infinite"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.002006,"raw_usage":{"total_tokens":7805,"prompt_tokens":902,"completion_tokens":6903,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":518,"completion_tokens_details":{"reasoning_tokens":6806}},"tokens_in":518,"tokens_out":6903,"duration_ms":51237,"temperature":1.0,"reasoning_tokens":6806,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T19:50:05.818882+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the orbit enumeration for the G₂ root lattice with its Weyl group acting on L/4L: if any W-stable subset of L/4L of size at most 5 generates the module, then the claimed lower bound c₂(k) ≥ 6 fails; the same check applies to A₃, F₄, D₅, and E₆ for the values 12, 24, 40, and 72.","supporting_citations":[],"review_version":1}