{"id":"097bf73d-c345-4d33-8091-02e9fa0e69de","arxiv_id":"1908.03105","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"The index of a type-D seaweed algebra is computed by counting meander components against a 'tail' set of vertices, yielding gcd-based closed forms and a Frobenius classification.","lead":"This paper gives formulas for the index of seaweed subalgebras of so(2n) by counting features of an associated planar graph called a meander. A generalist reader may care because index-zero (Frobenius) Lie algebras appear in deformation quantization and the classical Yang-Baxter equation.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 5.10's proof reduces to ∑a_i=n via Corollary 5.9, but that corollary is about ind, not the meander statistic 2C+P~; the central formula is left unproved for ∑a_i<n.","rationale":"I read Theorem 5.10 as the load-bearing result: it is what turns meander counts into index values, and Theorems 5.17–5.44 all assume it. The proof's reduction step is the weakest link because it asserts a sufficiency that Corollary 5.9 does not logically provide. The reader's Theorem 5.38 modulus concern is legitimate and checkable; I confirm the stated denominator n is inconsistent with the proof's modulo n/2 argument. But if Theorem 5.10's reduction gap hides a real counterexample, the entire edifice fails, whereas the 5.38 error is localized to one classification family. Hence I treat the proof gap as the primary concern while agreeing that the paper should be CONDITIONAL rather than rejected, since both issues are fixable by additional arguments or corrections.","tokens_in":22703,"tokens_out":21783,"duration_ms":207228,"concrete_test":"Enumerate all type-D seaweed-shaped pairs (a|b) with n≤10, including ∑a<n and t odd (configurations II/III), compute ind via the Panyushev–Yakimova reductions (Theorems 5.6–5.8), and independently count 2C+P~ from the meander. If every case matches, Theorem 5.10 survives and the missing f-transformation proof is merely expository; if any case with ∑a<n mismatches, the central formula must be revised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central index formula, Theorem 5.10, is proved only in the special case ∑a_i = n. The proof states 'By Corollary 5.9 and symmetry, it suffices to consider the case when ∑ ai = n and ∑ bi ≤ n,' but Corollary 5.9 is a statement about the index (ind pD_{n+k}(a|b) = k + ind pC_n(a|b) or k−1+...), not about the graph statistic f = 2C + P~ that Theorem 5.10 must show equals the index. For a seaweed with ∑a_i < n (tail configurations I and III of Definition 5.4), no argument is given that f(M^D_n(a|b)) equals the type-C statistic plus the appropriate constant. Thus the engine of all later closed-form formulas is unverified in a whole family of cases. This is a missing step rather than a demonstrated falsehood; the separate modulus error in Theorem 5.38 identified by the reader is real but downstream of this gap.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops meander-theoretic index formulas for seaweed subalgebras of so(2n), in parallel with earlier work in types A, B, and C. The authors introduce type-D meanders and a type-D tail with three configurations, prove a combinatorial formula for the index (Theorem 5.10), and use it to derive explicit gcd-type closed formulas and a classification of Frobenius type-D seaweeds, including a negative result in the spirit of Karnauhova-Liebscher. The paper also treats type-D seaweeds without seaweed shape and shows their indices differ from those of associated seaweed-shaped algebras by a constant.","tokens_in":22845,"tokens_out":16000,"duration_ms":175074,"significance":"If correct, the paper would supply a complete tail-based calculus for type-D seaweeds and would advance the program, initiated by Dergachev-Kirillov and Coll-Hyatt-Magnant, of computing indices of seaweed algebras from meander combinatorics in all classical types. The explicit gcd formulas and the type-D analogue of the no-polynomial-gcd obstruction are useful and natural contributions. However, the current version contains a false classification theorem and an unproved reduction in the proof of the central index formula, so the significance is conditional on substantial repair.","major_comments":[{"comment":"The stated condition (ii) is not the condition proved. The proof compares residues modulo n/2: after the display beginning 'Let σ1 ...', the congruences are modulo n/2, and the Euler argument yields k1 and k2 modulo n/2. But condition (ii) of the theorem evaluates the fractional part of (Δ/2)^(φ(n/2)-1)/n, which is a statement modulo n. These are not equivalent. For a concrete counterexample, take n=22, a=5, b=17, c=17. Then gcd(a+b,b+c)=gcd(22,34)=2, and Δ=a-c≡10 (mod 22), so Δ/2=5; moreover (5^9)/22 has fractional part 9/22≈0.409<0.5, so the theorem's hypotheses hold. Yet the residue modulo 11 is 9, which is larger than 11/4, and the meander has index 2 rather than 0. Thus Theorem 5.38 is false as stated. The denominator in (ii), and in the scholium following the proof, should be n/2, and the proof must also handle the excluded residue n/4 explicitly.","section":"Theorem 5.38 and its proof, Section 5.4.3"},{"comment":"The reduction to the case ∑ai=n is asserted but not proved. Corollary 5.9 is a statement about the index ind, not about the statistic f=2C+P~ used in Theorem 5.10. For a type-D meander with ∑ai<n (tail configurations I with t even, II, and III), no argument is given that f(M^D_n(a|b)) equals the type-C statistic on the reduced data plus the appropriate constant k or k−1. The induction that follows starts only after the reduction, in the case ∑ai=n. This leaves a genuine family of cases unverified in the proof of the central formula. The gap may be repairable, for instance by an explicit edge-contraction lemma in the spirit of Lemma 6.1, but the current text does not supply it.","section":"Theorem 5.10, proof, Section 5.4.1"}],"minor_comments":[{"comment":"The proof refers to 'equations (3) and (4)' when the displayed congruences are labeled (6) and (7).","section":"Proof of Theorem 5.34"},{"comment":"The scholium repeats the modulus error, writing φ(n) instead of φ(n/2) and dividing by n; it should be corrected together with Theorem 5.38.","section":"Scholium after Theorem 5.38"},{"comment":"There is a typo in the definition of t: the last summand is written as 'br' rather than 'bi', and the summation index should be made consistent.","section":"Corollary 5.9"},{"comment":"The word 'Futhermore' should be 'Furthermore'.","section":"Theorem 5.26"}],"recommendation":"major_revision","confidential_remarks":"The counterexample to Theorem 5.38 is a mathematical falsehood in a central theorem, and the gap in the proof of Theorem 5.10 affects the engine of the paper. I am not recommending rejection because both issues appear local and potentially repairable within the manuscript's framework: the modulus error can be fixed by changing the condition to the residue modulo n/2, and the reduction in Theorem 5.10 can likely be supplied by a separate contraction lemma. The authors should also re-check all downstream results that rely on Theorem 5.38."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nHere is my read. The paper is a real step forward for the meander-index program in type D. It gives a tail-based graph formula (Thm 5.10), works out the three tail configurations, and produces a battery of closed-form gcd formulas including Frobenius criteria. The no-polynomial-gcd theorem, if made rigorous, would be a satisfying type-D analogue of Karnauhova-Liebscher. The section on type-D seaweeds that do not have seaweed shape (Thm 5.1, Thm 5.42) is genuinely new and useful.\n\nThe good news: Theorem 5.10 is proved by an induction that is mostly coherent, and the surrounding graph machinery is well organized. The authors are honest about relying on Panyushev-Yakimova and Dvorsky for the inductive index reductions; self-citation is not a problem here.\n\nThe soft spots are real. First, the proof of Thm 5.10 says that by Cor 5.9 it suffices to treat the case sum a_i = n. But Cor 5.9 is an index identity, not a statement about the graph statistic f = 2C + P~. Unless you add a lemma showing f(M^D_{n+k}(a|b)) = k + (or k-1+) f(M^C_n(a|b)), the formula is only verified in the sum a_i = n case. I do not think the gap is fatal, but it is a genuine missing step.\n\nSecond, Thm 5.38 as stated is false. The theorem compares the fractional part of (Delta/2)^(phi(n/2)-1)/n to 1/2, but the proof compares residues modulo n/2. Those two conditions are not equivalent. The reader's example n=22, a=5, b=17, c=17 satisfies the stated gcd and fractional-part conditions while the seaweed has index 2. This is a concrete counterexample, not a stylistic slip. The criterion can probably be repaired by replacing n with n/2 in the denominator, but as written it cannot stand.\n\nThird, Thm 5.40(ii) is argued by informal reduction to type A. The intuition is plausible, but the proof is a sketch. If the theorem is important to the paper's scope, it needs a real transfer argument.\n\nNet: the paper deserves a serious referee. The main framework is sound and the errors are local, but a referee should demand a fixed Thm 5.38, a proof of the missing comparison lemma in Thm 5.10, and a rigorous version of Thm 5.40(ii) before the classification results are quoted. I would not cite the Frobenius classification as it stands.","headline":"Valuable type-D meander program with a repairable proof gap and a false Frobenius criterion; deserves refereeing but needs a major revision.","tokens_in":23404,"tokens_out":6844,"would_cite":false,"duration_ms":73941,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["17B08"],"pacs":[],"model":"deepseek-v4-flash","headline":"For a type-D seaweed algebra, the index equals 2C + P-tilde, where C counts cycles and P-tilde counts tail-meeting paths in the meander.","keywords":["Frobenius Lie algebra","special orthogonal Lie algebra","seaweed","index","meander","tail","greatest common divisor formula","type D"],"falsifier":"Take the paper's own example $p^D_{14}((5|9)/(9), III)$, whose meander has index zero. With the representative 5 for $\\Delta/2 \\bmod 7$, the fractional part in Theorem 5.38 is $5^5/14 \\approx 0.214$, matching the Frobenius verdict; the congruent representative 12 gives $12^5/14 \\approx 0.714$, which would declare the same algebra non-Frobenius. Any convention that fixes representatives must choose the first value, and a direct index calculation on any example where the two representatives straddle 1/2 settles whether the stated condition is correct.","tokens_in":22436,"feed_emoji":"🌿","tokens_out":15724,"duration_ms":144124,"temperature":0.7,"pith_summary":"This paper completes a program that computes the index of seaweed Lie algebras by counting pieces of an associated planar graph, now for the type-D family $so(2n)$. It proves that for a seaweed defined by two partial compositions $a$ and $b$, the index is $2C + \\tilde P$, where $C$ is the number of cycles in the meander and $\\tilde P$ is the number of path components that meet the distinguished tail in zero or two vertices. From this single formula the paper extracts greatest-common-divisor index formulas for low-complexity seaweeds, characterizes Frobenius type-D seaweeds, and shows exactly where closed-form gcd formulas stop being possible. A reader should care because the same meander machinery that worked in types A and C now covers the special orthogonal case, making the index an elementary counting problem rather than a computation in the dual of a Lie algebra.","feed_headline":"Type-D seaweed index: count meander cycles and tail paths","feed_subtitle":"The same meander trick used for types A and C now yields gcd and Frobenius tests for so(2n).","key_machinery":"The load-bearing object is the meander, a planar graph whose top edges are laid out according to one composition and bottom edges according to the other, so that each vertex is incident with at most one edge of each kind and the graph decomposes into cycles and paths. For type D the paper adds a distinguished set of vertices, the tail $T_n^D(a|b)$, defined by the difference between the two composition lengths and taking one of three configurations depending on parity and on whether the first composition sums to $n$. The index formulas count cycles and tail-meeting paths of this graph, and the closed-form results additionally use the winding-down moves and homotopy types that reduce a meander to a sequence of component eliminations, together with the observation that in one-parameter cases the relevant permutation is generated by a single difference $\\Delta$.","core_discovery":"The paper's central discovery is a graph-theoretic formula for the index of a type-D seaweed subalgebra of $so(2n)$, stated as Theorem 5.10: if $M_n^D(a|b)$ is the meander built from the two partial compositions $a$ and $b$, and $T$ is the associated tail, then $\\operatorname{ind} p_n^D(a|b) = 2C + \\tilde P$, where $C$ is the number of cycles in the meander and $\\tilde P$ is the number of path components containing either zero or two vertices of $T$. Equivalently, the index counts cycles of the top-bottom permutation containing zero or two tail elements. From this base formula the paper derives explicit greatest-common-divisor and congruence tests for Frobenius seaweeds, including the one-part closed forms and the three-part classifications, and proves that four-part seaweeds with a proper second parabolic cannot have their index given by any polynomial gcd formula. It also reduces non-seaweed-shaped type-D seaweeds to seaweed-shaped ones, with the index changing by zero or two, and identifies Frobenius examples among them with type-A meanders of homotopy type $H(2)$.","pith_inferences":["The fractional-part tests in the paper can be read as a comparison of two modular inverses; this suggests they are checkable by fast modular exponentiation once a representative of $\\Delta/2$ is fixed, giving a polynomial-time Frobenius test for one-parameter type-D seaweeds.","The standard reduction from type C to type B suggests that the same tail formula should yield gcd and congruence classifications for the odd orthogonal family $so(2n+1)$, which the paper does not tabulate explicitly.","The switch relating non-seaweed-shaped and seaweed-shaped algebras points to an algorithm that needs only the Dynkin subsets, not the matrix realization: replace the exceptional root, compute the seaweed-shaped index, and adjust by 0 or 2 according to whether two specified vertices lie on a path or a cycle."],"forward_implications":["A type-D seaweed is Frobenius exactly when its meander is a forest rooted in the tail, so index-zero algebras can be recognized by looking at the picture.","Tail configuration II can never be Frobenius, so the search for Frobenius algebras reduces to configurations I and III.","For one-part and three-part seaweeds, index and Frobenius status are decided by gcds and congruence or fractional-part conditions on the part sizes.","Four-part seaweeds with a proper second parabolic admit no polynomial gcd index formula, so the list of closed forms is complete at four parts.","Seaweeds without seaweed shape have index equal to that of a seaweed-shaped cousin, or that value minus two, and the Frobenius ones correspond exactly to type-A meanders of homotopy type $H(2)$."],"supporting_citations":[{"why":"It introduces meanders and connects their graph components to the index of seaweed subalgebras in type A, the template every later formula extends.","marker":"[6]"},{"why":"It supplies the inductive index formulas for seaweeds, including the type-D reductions used in the proof of Theorem 5.10.","marker":"[14]"},{"why":"It develops type-D meanders and supplies the corollary used to compute indices of seaweeds without seaweed shape.","marker":"[16]"},{"why":"It defines symplectic meanders and the type-C tail formula that the type-D tail and $2C+\\tilde P$ formula generalize.","marker":"[4]"},{"why":"It gives the parabolic index base cases for $so(2n)$ used in the tail-configuration arguments.","marker":"[8]"},{"why":"It proves the type-A no-polynomial-gcd obstruction that the type-D negative result in Theorem 5.40 is modeled on.","marker":"[13]"},{"why":"It defines the winding-down moves, signatures, and homotopy types $H(k)$ used in the three-part and Frobenius analyses.","marker":"[2]"}],"fun_headline_variants":["Type-D seaweed index: count meander cycles and tail paths","Index of type-D seaweeds: meander cycles and tails","Seaweed index formulas for type-D via meander graph","Type-D seaweed index: tail paths and cycles give gcd"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The Frobenius classification for the four-tail case in Theorem 5.38 depends on treating $\\Delta/2$, which is only defined modulo $n/2$, as a concrete integer when forming the fractional part in the theorem's condition; different representatives of the same residue class can change the verdict.","fun_headline_variants_meta":{"raw":{"variants":["Type-D seaweed index: count meander cycles and tail paths","Index of type-D seaweeds: meander cycles and tails","Seaweed index formulas for type-D via meander graph","Type-D seaweed index: tail paths and cycles give gcd"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00022,"raw_usage":{"total_tokens":1425,"prompt_tokens":900,"completion_tokens":525,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":516,"completion_tokens_details":{"reasoning_tokens":455}},"tokens_in":516,"tokens_out":525,"duration_ms":5882,"temperature":1.0,"reasoning_tokens":455,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:32:25.455236+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the paper's own example $p^D_{14}((5|9)/(9), III)$, whose meander has index zero. With the representative 5 for $\\Delta/2 \\bmod 7$, the fractional part in Theorem 5.38 is $5^5/14 \\approx 0.214$, matching the Frobenius verdict; the congruent representative 12 gives $12^5/14 \\approx 0.714$, which would declare the same algebra non-Frobenius. Any convention that fixes representatives must choose the first value, and a direct index calculation on any example where the two representatives straddle 1/2 settles whether the stated condition is correct.","supporting_citations":[{"cited_title":"Dergachev and A","cited_arxiv_id":null,"evidence_quote":"It introduces meanders and connects their graph components to the index of seaweed subalgebras in type A, the template every later formula extends."},{"cited_title":"Panyushev","cited_arxiv_id":null,"evidence_quote":"It supplies the inductive index formulas for seaweeds, including the type-D reductions used in the proof of Theorem 5.10."},{"cited_title":"Panyushev and O","cited_arxiv_id":null,"evidence_quote":"It develops type-D meanders and supplies the corollary used to compute indices of seaweeds without seaweed shape."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It defines symplectic meanders and the type-C tail formula that the type-D tail and $2C+\\tilde P$ formula generalize."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It gives the parabolic index base cases for $so(2n)$ used in the tail-configuration arguments."},{"cited_title":"Connected components of meanders: I. Bi-rainbow meanders","cited_arxiv_id":"1504.03099","evidence_quote":"It proves the type-A no-polynomial-gcd obstruction that the type-D negative result in Theorem 5.40 is modeled on."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It defines the winding-down moves, signatures, and homotopy types $H(k)$ used in the three-part and Frobenius analyses."}],"review_version":1}