{"id":"d8b49615-d398-4bea-9706-c2ecce49302e","arxiv_id":"2602.14220","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"On every almost abelian Lie algebra, presymplectic forms of rank R exist exactly when a rank formula in the Jordan data of ad_e holds, and symplectic forms form a finite moduli space, each class a permuted canonical form.","lead":"Complete existence criterion for left-invariant closed 2-forms of every rank on all almost abelian Lie algebras, read off the Jordan form of ad_e; plus a rigidity theorem: the moduli space of symplectic forms is finite, every class being a permuted canonical form up to automorphism and scaling. Read it for a full classification answer in solvable Lie-group geometry, and for reusable canonical representatives of a natural matrix congruence.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Prop 4.22's 'apply Corollary 4.10/4.12 to all rows' step is not justified: mixed diagonal/off-diagonal configurations can violate the lemmas' hypotheses, so the canonical-form proof has a genuine gap.","rationale":"The reader's weakest_assumption correctly identifies the completeness of the block-by-block normalization in Prop 4.22/5.18 as the pivotal gap. My stress-test confirms this is not a merely cosmetic issue: there exist valid maximal-rank matrices in ˙H^4 whose nonzero blocks are arranged so that none of the normalization lemmas apply directly, and the proof gives no algorithm for reducing them. Since Thm 6.3's proof depends entirely on these propositions, this is the single most load-bearing concern. I do not see a counterexample to the theorem itself; the conclusion may still be true, and the gap may be fillable with a more careful simultaneous-normalization argument. The other errors noted by the reader (strict inequality in Prop 5.15, R_JNR=1 in Cor 6.2's proof) are mechanical typos that do not affect the structural concern. Therefore the reader's CONDITIONAL verdict remains appropriate, and no verdict change is needed.","tokens_in":35360,"tokens_out":26586,"duration_ms":218373,"concrete_test":"For J_NR = J_2(0)⊕J_2(0)⊕J_1(0), define the 5×5 matrix C with 2×2 blocks C11 = C12 = C21 = [[1,0],[0,−1]], C22 = 0, and zero third row/column. Verify C ∈ ˙H^4 and rank(C)=4. Then solve symbolically for S ∈ ˙T_JNR such that S* C S = P_NR P^t J(5,4) P, where P is a permutation matrix. If no solution exists, Prop 4.22 is false; if a solution exists, extract the construction and use it to fill the simultaneous-normalization gap. Repeat the analogous test for the complex case of Prop 5.18 using J_NC = C_1(0,b)⊕C_1(0,b) with a mixed diagonal/off-diagonal configuration.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central finiteness theorem (Thm 6.3) rests entirely on Prop 4.22 and 5.18. Their proofs assert that one can 'apply Corollary 4.10 or Corollary 4.12 to all the rows and the corresponding columns' and obtain a direct sum of normalized blocks. But the lemmas have hypotheses that are not shown to hold simultaneously. In particular, Lemma 4.9 (diagonal normalization) requires c^(q)_il = c^(q)_li = 0 for all same-size i, and Lemma 4.11 (off-diagonal normalization) requires c^(q)_ll = c^(q)_mm = 0 for q ≤ k. A matrix can have a row where both a diagonal block and an off-diagonal block are nonzero, with the partner row having only the off-diagonal block. Such a configuration satisfies none of the lemma hypotheses, yet can occur at maximal rank. Concretely, for J_NR = J_2(0)⊕J_2(0)⊕J_1(0), take the 5×5 matrix C with 2×2 blocks C11 = C12 = C21 = [[1,0],[0,−1]], C22 = 0, and the third row/column zero. This C lies in ˙H^4 and has rank 4, but Cor 4.10 fails for row 1 (because C21 ≠ 0), Cor 4.12 fails for the pair (1,2) (because C11 ≠ 0), and Lemma 4.11 fails for the same reason. The proof does not explain how to proceed. The identical gap appears in Prop 5.18 via Cor 5.9/5.11. If no S ∈ ˙T_JNR conjugates this C to the claimed canonical form, Prop 4.22 is false and Thm 6.3 loses its proof; if such S exists, the missing simultaneous normalization step still needs to be supplied.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies left-invariant closed 2-forms of fixed rank on almost abelian Lie algebras. It reduces the existence and equivalence problems to a matrix congruence problem for solutions B of BJ_N + J_N^t B = 0 modulo the centralizer of J_N. The first main result, Theorem 6.1, gives an explicit rank criterion in terms of two numerical invariants R_{J_{N_R}} and R_{J_{N_C}}. The second main result, Theorem 6.3, claims that for the maximal (symplectic) rank the moduli space PΩ²_{D,closed}(g) is finite and that every class is represented by a permutation of the canonical symplectic form. The matrix-theoretic core is Propositions 4.22 and 5.18, which purport to provide canonical forms for maximal-rank solutions by a block-elimination algorithm.","tokens_in":35663,"tokens_out":9365,"duration_ms":87071,"significance":"If correct, these are strong and useful results: a complete existence criterion for all presymplectic ranks on all almost abelian Lie algebras, and a very rigid description of the symplectic moduli space. The reduction to the 'dotted' congruence quotient and the reliance on standard matrix-theoretic tools are appealing, and Theorem 6.1 is explicit enough to be applied directly. However, the proof of the canonical-form propositions that support Theorem 6.3 contains a substantial gap, so the significance of the paper is conditional at this stage. The potential contribution is real: the paper identifies the right matrix problem and gives a credible route to the answer.","major_comments":[{"comment":"The proof (p. 25) says 'We can then apply Corollary 4.10 or Corollary 4.12 to all the rows and the corresponding columns.' This simultaneous normalization is not justified. Corollary 4.10 is derived from Lemma 4.9, whose hypotheses require that when C_ll has first nonzero diagonal at k, all same-size off-diagonal entries in row/column l have zero diagonals up to k. Corollary 4.12 requires C_ll = C_mm = 0. A maximal-rank matrix can contain a row with both a nonzero diagonal block and a nonzero same-size off-diagonal block, satisfying none of these hypotheses. For J_{N_R}=J_2(0)⊕J_2(0)⊕J_1(0), take the 5×5 block matrix with C_11=C_12=C_21=diag(1,-1), C_22=0, and the third row/column zero. This matrix lies in ˙H^4_{J_{N_R}} and has rank 4, but Corollary 4.10 fails for row 1 (C_12≠0), Corollary 4.12 fails for the pair (1,2) (C_11≠0), and Lemma 4.11 also fails. The proof gives no alternative.","section":"§4.2.2, Proposition 4.22"},{"comment":"The proof (p. 38) makes the same assertion: 'we can apply Corollary 5.9 or Corollary 5.11 to all rows and corresponding columns.' The complex variants inherit the same simultaneous-hypothesis conditions, and the proof refers back to the real case for details, so the gap is not repaired. A correct proof would need to specify an order of elimination that either keeps previously normalized blocks intact or shows that the hypotheses of the applicable lemma are satisfied at each step; no such argument is present.","section":"§5.3.2, Proposition 5.18"},{"comment":"The proof (p. 39) contains the sentence 'Define Y=R R and W=R−2−Y', which appears to be a typo; if Y=R then W=−2. The proof also needs to state explicitly how the nonemptiness of H^{R−2}_{J_N} yields a closed 2-form of rank R. Proposition 3.7 contains the lifting step (choose v outside Im B), but the proof of Theorem 6.1 should either invoke it or fill in the missing rank-arithmetic. This is a localized defect, but it concerns the proof of the first main theorem.","section":"§6, Theorem 6.1"}],"minor_comments":[{"comment":"The congruence '= N mod 2' uses N, but the discussion concerns N_R; it should presumably be N_R mod 2. Please clarify.","section":"§4.2, Eq. (4.44)"},{"comment":"The statement says a matrix 'commutes with J_{N_C}', but the displayed equation is BJ_{N_C}+J_{N_C}^t B=0. The wording should be 'satisfies (5.18)' to avoid confusion.","section":"§5.2, Proposition 5.5"},{"comment":"The displayed matrix has an entry α in the (1,1) position, but the condition is written as AJ_N − J_N A = 0 with no α. If α is intended to be absorbed into A or into the scaling action, this should be stated explicitly.","section":"§3, Corollary 3.5"},{"comment":"There are several typographical slips (e.g., 'Fist' in the introduction, 'we can apply can' in the proof of Proposition 4.22, and 'by joining this results' in the proof of Theorem 6.3). These should be corrected in a revision.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The gap identified in Propositions 4.22 and 5.18 is genuine and load-bearing for Theorem 6.3. I do not see a counterexample to the theorem itself, and the proposed canonical forms may well be correct, so rejection seems premature. The paper would be acceptable if the authors supply a correct simultaneous-normalization proof or an alternative induction for the maximal-rank case, and if they repair the proof of Theorem 6.1. The external tools cited are appropriate, and I did not find circularity."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a real contribution, and the two headline theorems are probably true. But the proof of the central normalization step (Prop 4.22 and its complex twin 5.18) is incomplete as written, and a referee should demand a repaired argument before acceptance.\n\nWhat's genuinely new: Thm 6.1 gives a complete existence criterion for presymplectic forms of every rank on all almost abelian Lie algebras, in terms of the Jordan data R_JNR and R_JNC. I checked the rank-2 case and the reduction to the centralizer subgroup (Cor 3.5); those work. Thm 6.3 — finiteness of the symplectic moduli space plus the statement that every class is a permuted canonical form — is a strong, clean result if the matrix machinery holds. The paper is honest about the limits of its method in Remark 4.13 and cites the relevant literature appropriately.\n\nThe soft spot is exactly where the reader put it. In Prop 4.22, the proof says to 'apply Corollary 4.10 or Corollary 4.12 to all the rows and the corresponding columns.' That is not justified. The stress-test example is a good one: take J_NR = J_2(0)⊕J_2(0)⊕J_1(0), and C with 2×2 blocks C11 = C12 = C21 = [[1,0],[0,−1]], C22=0, last row/col zero. This C is in ˙H^4 and has rank 4. For row 1, Lemma 4.9's hypothesis fails because C21≠0; Lemma 4.11's hypothesis fails because C11≠0. So neither corollary applies, and the proof gives no way to proceed. The same gap appears in Prop 5.18 via Cor 5.9/5.11. This is a real hole in the proof of Thm 6.3. I don't have evidence the theorem itself is false — the example may still be congruent to canonical form via a more careful argument — but the paper as written does not supply that argument.\n\nThere are also a few small misstatements: Prop 5.15's strict inequality should be ≤, and Cor 6.2's proof sets R_JNR=1 even though R_JNR is always even; the intended conclusion is R_JNR=0. These are minor and fixable.\n\nBottom line: worth serious refereeing. The existence criterion alone justifies it. Send to a differential-geometry or matrix-theory referee who is willing to check the normalization lemma carefully, and ask for a complete proof of Props 4.22/5.18.","headline":"Genuinely new results, mostly sound, but the proof of the key canonical-form proposition has a real gap — needs revision, not rejection.","tokens_in":36340,"tokens_out":7148,"would_cite":true,"duration_ms":61540,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["17B30","15A21","53D05","22E25"],"pacs":[],"model":"deepseek-v4-flash","headline":"On every almost abelian Lie algebra, the moduli space of symplectic forms is finite and every symplectic form is a permuted canonical 2-form; rank-R presymplectic forms obey an explicit eigenvalue-pairing criterion.","keywords":["presymplectic forms","almost abelian Lie algebras","symplectic moduli space","left-invariant 2-forms","real Jordan normal form","matrix congruence","permutation normal form","Lie groups"],"falsifier":"For J_5 = J_3(λ) ⊕ J_2(-λ), the five-dimensional real case of Example 4.21, take the maximal-rank solution A_1 displayed there and compute the full congruence orbit {A^t A_1 A : A in T_{J_5}}. Proposition 4.22 predicts every orbit member is congruent to P^t J_{5,4} P for some permutation P. If a computer search finds an orbit element whose support cannot be reduced to that permutation pattern by any A in T_{J_5}, the canonical-form lemma — and with it Theorem 6.3 — is false.","tokens_in":35070,"feed_emoji":"🔄","tokens_out":10962,"duration_ms":87257,"temperature":0.7,"pith_summary":"Almost abelian Lie algebras are built from a codimension-one abelian ideal and one extra generator whose bracket acts by a linear map T. The paper asks which ranks of closed, nondegenerate left-invariant 2-forms these algebras admit, and how many such forms exist up to automorphism and scale. It answers the first question completely: a rank-R presymplectic form exists exactly when R is at most 2 plus a number computed from the real Jordan form of T, counting how many positive and negative eigenvalue blocks can be paired and how odd-sized zero blocks are left over. It answers the second question for maximal rank: the moduli space of symplectic forms is finite, and each class contains the canonical symplectic form with its coordinate labels permuted. The reason to care is that an infinite moduli problem collapses to a finite permutation count, with an explicit normal form usable for computation.","feed_headline":"Finitely many symplectic classes on any almost abelian Lie algebra","feed_subtitle":"The paper reduces the moduli problem to a finite permutation count from the real Jordan form.","key_machinery":"The central object is the matrix congruence action on the solution space H^K_{J_N} = {B skew-symmetric : B J_N + J_N^t B = 0, rank K}, modulo congruence by A in T_{J_N} = {A : A J_N = J_N A}. On this space the paper builds a block calculus: the solutions have upper-alternating-Toeplitz or lower-alternating-Hankel blocks, and a sequence of explicit congruences (Lemmas 4.8–4.14, with complex analogue 5.7–5.11) isolates each nonzero block as a standard block I^{k±} and kills the corresponding row and column. Proposition 4.22 and Proposition 5.18 assert that at maximal rank this process ends in a direct sum of standard blocks, i.e. a permuted copy of the canonical matrix J. The rank criterion us","core_discovery":"The paper's central claims are Theorem 6.1 and Theorem 6.3. Theorem 6.1 states that, writing ad_{e_1} in real Jordan normal form as a direct sum of a real part J_{N_R} and a complex part J_{N_C}, a rank-R presymplectic form exists if and only if R ≤ 2 + R_{J_{N_R}} + R_{J_{N_C}}, where each R_J is N (or N_C) minus the number of unmatched odd-sized zero Jordan blocks minus the Manhattan distance between the ordered lists of block sizes for λ and −λ. Theorem 6.3 states that PΩ²_{D,closed}(g) is finite and that every class [ω] satisfies [ω] = [P·ω_0] for a permutation P fixing e_1. In other words, every left-invariant symplectic form on an almost abelian Lie algebra is, up to automorphism and n","pith_inferences":["Editorial inference: the same permutation description may hold for rank-R presymplectic forms for every R once the non-maximal normalization gaps flagged in Remark 4.13 are resolved; Lemma 4.16 shows ranks descend by shifting blocks, which is the natural route.","Editorial inference: because the equivalence relation includes scale, the result is really about conformal symplectic classes; dropping scale would likely introduce additional discrete invariants, such as ratios of Jordan block sizes within a paired λ/−λ family.","Editorial inference: the finite moduli space suggests that geometric invariants of almost abelian symplectic Lie groups — for instance compatible complex or Kähler structures — may also admit finite or combinatorial classifications depending only on the Jordan form and a permutation.","Editorial inference: a direct count of the classes is not given, but the normal form reduces the problem to computing the stabilizer of P·ω_0 in T_{J_N}; that computation would turn finiteness into an exact cardinality."],"forward_implications":["Existence of presymplectic forms of any rank on any almost abelian Lie algebra is decidable by a finite computation from the real Jordan normal form of the defining map.","The moduli space of left-invariant symplectic forms on any almost abelian Lie algebra is a finite set; there are no continuous families of inequivalent symplectic structures.","Every symplectic form has a canonical representative P·ω_0, so invariant quantities of any symplectic form reduce to invariants of the permuted standard form.","The maximal-rank matrix congruence normal form is a standalone statement: skew-symmetric solutions of the Lyapunov-type equation with maximal rank are classified, up to the relevant congruence, by permutations.","The paired-eigenvalue conditions of Corollary 6.2 give an explicit list of which almost abelian Lie algebras admit symplectic forms at all."],"fun_headline_variants":["Symplectic moduli finite on almost abelian Lie algebras","All symplectic forms permute one canonical 2-form","Rank conditions for presymplectic forms from Jordan data","Finite symplectic classes: every one is a permutation","Presymplectic existence settled by Jordan block matching"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The argument for the finite moduli space assumes that the row-and-column cleaning used in Propositions 4.22 and 5.18 can be applied to every nonzero block in any order, with no later cleaning step undoing the normalization of an earlier one; if that simultaneous compatibility fails, the permutation normal form is not established.","fun_headline_variants_meta":{"raw":{"variants":["Symplectic moduli finite on almost abelian Lie algebras","All symplectic forms permute one canonical 2-form","Rank conditions for presymplectic forms from Jordan data","Finite symplectic classes: every one is a permutation","Presymplectic existence settled by Jordan block matching"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000227,"raw_usage":{"total_tokens":1293,"prompt_tokens":711,"completion_tokens":582,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":455,"completion_tokens_details":{"reasoning_tokens":499}},"tokens_in":455,"tokens_out":582,"duration_ms":5288,"temperature":1.0,"reasoning_tokens":499,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T23:21:06.058152+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For J_5 = J_3(λ) ⊕ J_2(-λ), the five-dimensional real case of Example 4.21, take the maximal-rank solution A_1 displayed there and compute the full congruence orbit {A^t A_1 A : A in T_{J_5}}. Proposition 4.22 predicts every orbit member is congruent to P^t J_{5,4} P for some permutation P. If a computer search finds an orbit element whose support cannot be reduced to that permutation pattern by any A in T_{J_5}, the canonical-form lemma — and with it Theorem 6.3 — is false.","supporting_citations":[],"review_version":1}