{"id":"3c112b81-1e7b-4626-bdd8-ea4da42abf79","arxiv_id":"2505.04581","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Borel orbits in the AIII and CII symmetric varieties are in bijection with matchings in (double) corona graphs, and this dictionary proves the Can-Ugurlu conjecture about non-integral orbit-counting coefficients.","lead":"This paper converts the problem of counting Borel orbits in three families of symmetric spaces into the much simpler problem of counting matchings in small graphs, and uses that dictionary to settle an open conjecture about the number of orbits in one of the families. The new combinatorial model offers cleaner proofs and new numerical results across classical symmetric varieties.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Type BI parametrization (Theorem 5.16) is asserted without proof; since Theorem 6.4 depends on it and also uses asymptotic O-bounds to infer a leading coefficient, the Can-Ugurlu conjecture is not established as written.","rationale":"The reader's weakest_assumption correctly identifies Theorem 5.16 as the most load-bearing gap: the Can-Ugurlu conjecture proof in Theorem 6.4 directly invokes this asserted parametrization, and the carry-over sentence in Section 5.3 is not a proof. The symplectic case itself contains an explicit omitted proof (Proposition 5.10), so the orthogonal case, which is only 'with minimal modifications,' is doubly unsupported. I also flag a secondary gap in Theorem 6.4: the asymptotic reasoning bounds b_{m,n} by O(m^{2n+1}S_n) and then concludes the leading coefficient equals S_n, which is not a valid inference from an upper bound. That step is likely repairable by writing exact leading coefficients of the subset counts, but as written it is an additional unproven assertion. The AIII part is proven in detail, so the paper's overall conditional status is appropriate. My proposed computational tests would provide direct evidence for or against the missing BI parametrization and the claimed leading coefficients, making the conditional more concrete.","tokens_in":12884,"tokens_out":13919,"duration_ms":121357,"concrete_test":"Enumerate type BI Borel orbits for small (m,n) via Wyser's clan parametrization (arXiv:1201.4397) or Can-Ugurlu's explicit polynomials, and compare with the number of minus-invariant 2m-matchings in C_{2m+2n+1} for (m,n)=(1,1),(1,2),(2,1). Agreement would support the missing Theorem 5.16; disagreement would falsify it. Separately, compute the exact polynomial b_{m,1} and b_{m,2} (e.g., from Can-Ugurlu's formulas) and check their leading coefficients are 7/6 and 27/40; if not, the asymptotic derivation's inferred leading coefficient is wrong.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 5.3 states 'The arguments in the previous subsection carry over to the case of orthogonal groups with minimal modifications' but gives no orthogonal analogue of Lemma 5.9, Proposition 5.10, or the single-double-coset step. Moreover, even the symplectic Proposition 5.10 has its proof omitted ('we will omit the proof'). Theorem 6.4 (Can-Ugurlu conjecture) rests directly on Theorem 5.16: if the missing arguments fail, b_{m,n} is not the number of minus-invariant (2n+1)-matchings. Independently, the proof of Theorem 6.4 uses asymptotic O-estimates: it shows b_{m,n}=O(m^{2n+1}S_n) with S_n<1 and then concludes the leading coefficient is S_n, but an upper bound does not determine a leading coefficient; the matching-count leading term must be shown to equal the all-subsets leading term, not merely bounded by it. The AIII bijection (Section 4) is proven; the failure mode is localized to the unproven BI parametrization and the non-rigorous leading-coefficient step.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper proposes alternative combinatorial parametrizations of Borel orbits in classical symmetric varieties by matchings of corona graphs. Theorem 1.2 asserts a natural bijection for type AIII, (GL_{m+n}, GL_m × GL_n), between B(k)-orbits in X(k) and m-matchings of the corona graph C_{m+n}, and, for char(k) ≠ 2, a similar bijection for type CII, (Sp_{2m+2n}, Sp_{2m} × Sp_{2n}), using the double corona graph. The paper also states a type BI parametrization (Theorem 5.16) for SO_{2m+2n+1}/S(O_{2m} × O_{2n+1}) and uses it to prove the Can–Uğurlu conjecture (Theorem 6.4) that the Borel-orbit-counting polynomial b_{m,n} has degree 2n+1 and non-integral leading coefficient for n>0. Additional applications include recurrences, unimodality, and ultra log-concavity for the AIII and CII orbit counts.","tokens_in":12959,"tokens_out":21846,"duration_ms":222495,"significance":"If the proofs are completed, the matching model would be a clean and useful alternative to clan parametrizations, and the confirmation of the Can–Uğurlu conjecture would be a genuine result. The AIII path is written out in reasonable detail and the application of matching-polynomial results to obtain unimodality and log-concavity is elegant. However, the two most important claims—the type BI parametrization and the non-integrality theorem for its orbit-count polynomial—are not established as written: one is asserted without proof, and the other relies on an invalid asymptotic inference. The CII parametrization also has a substantial omitted proof at a load-bearing point.","major_comments":[{"comment":"Theorem 5.16, the parametrization of Borel orbits for type BI by minus-invariant matchings, is asserted without proof. The sentence 'The arguments in the previous subsection carry over to the case of orthogonal groups with minimal modifications' is not a proof. No orthogonal analogue of Lemma 5.9, Proposition 5.10, or Proposition 5.14 is supplied, and Remark 5.17(2) itself notes a genuine difference from the symplectic case. Since Theorem 6.4 and the claimed confirmation of the Can–Uğurlu conjecture depend directly on Theorem 5.16, this omission is load-bearing.","section":"§5.3, Theorem 5.16"},{"comment":"The sufficiency direction for the type CII bijection is left with 'we will omit the proof.' Proposition 5.10 is the step that shows a minus-invariant matching with no horizontal edges supports a k-point of G^φ, and without it Theorem 1.2(2) and all applications involving c_{p,m} in Section 6 rest on an unproved claim. Additionally, Proposition 5.14 is proved only under the hypothesis k = \\bar{k}, while Theorem 5.15 is stated for double cosets over a general field; the paragraph following Theorem 5.15 about replacing k by \\bar{k} does not supply a proof for arbitrary k, it only explains why base change to the algebraic closure fails.","section":"§5.2, Proposition 5.10 and Theorem 5.15"},{"comment":"The proof of non-integrality of the leading coefficient uses only upper bounds. The displayed estimate b_{m,n} = O( m^{2n+1} ∑_{l=0}^n 1/((2l+1)!(n-l)!) ) with the sum less than 1 does not determine the leading coefficient: an O-bound carries an arbitrary multiplicative constant, so the actual coefficient could be larger than the displayed sum. One needs an asymptotic equivalence, or an explicit upper bound with constant 1, together with a separate lower bound showing the degree is exactly 2n+1. As written, the leading coefficient of b_{m,n} is not computed or bounded in the required way.","section":"§6, Theorem 6.4"},{"comment":"Theorem 3.4 is a Gabriel-type bijection asserted for arbitrary field k and quoted from [4]. Classical Gabriel's theorem is usually stated over algebraically closed fields, and it is not immediate that indecomposable representations of Q_p over a non-algebraically-closed field are classified by positive roots without additional hypotheses. Because Theorem 1.2(1) is claimed for arbitrary k, the precise statement and reference, or a proof, should be supplied so that the AIII bijection is not resting on an unverified variant of Gabriel's theorem.","section":"§3, Theorem 3.4"}],"minor_comments":[{"comment":"The coloring rule in Theorem 4.3 does not specify what to do with admissible roots of the form L_i + L_{m+n+2}. Example 4.6 shows that such roots appear as auxiliary summands that are not colored, and this is essential for the number of colored edges to equal m. Please state this explicitly in the proof of Theorem 4.3.","section":"§4, Theorem 4.3 and Example 4.6"},{"comment":"Proposition 1.3 states the ultra-log-concavity inequality with a factor (1 + 1/n), while Proposition 6.2 states the same inequality with (1 + 1/m). Since n is not defined in Proposition 1.3, this appears to be a typo and should be corrected.","section":"§1, Proposition 1.3 vs. §6, Proposition 6.2"},{"comment":"The proof of Proposition 5.7 refers to 'some suitable choices of the signs ±' but does not specify how the signs are chosen or why they do not affect the double coset. A short clarification would improve readability.","section":"§5.2, proof of Proposition 5.7"},{"comment":"The term 'm-matching' is used throughout without a formal definition in Section 1 or Section 2. Given that the paper's main results rest on this graph-theoretic notion, a definition should be included.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The manuscript contains two substantial unproved planks—Theorem 5.16 and Proposition 5.10—and the proof of the headline conjecture has a genuine asymptotic gap. These are fixable within the scope of the paper if the author supplies full proofs, but I would not accept the paper in its current form. I also encourage the editor to ensure that the quiver-theoretic input for arbitrary fields (Theorem 3.4) is verified by a specialist, since it is cited from a paper on canonical bases rather than from a standard Gabriel-theorem reference."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The real news here is the AIII matching parametrization. The proof through complementary quiver representations is written out in enough detail to be checked, and it gives a genuinely new dictionary: Borel orbits for GL_{m+n}/GL_m×GL_n correspond to m-matchings in the corona graph. The binary-matrix representatives, the duality operation, and the recovery of the known recurrences are all clean. That part is the paper's contribution and it holds up.\n\nThe CII part is a different story. The minus-action on matchings is described explicitly, and Lemma 5.9 is proved, but the sufficiency direction (Proposition 5.10) is explicitly left with “we will omit the proof.” That is a load-bearing omission. Section 5.3 then asserts Theorem 5.16 for the orthogonal group by saying the arguments carry over with minimal modifications, without giving the orthogonal analogues of the key lemmas. Theorem 5.16 is not a side remark: the proof of the Can–Ugurlu conjecture in Theorem 6.4 rests directly on it. So the headline conjecture is not established as written.\n\nThere is also a smaller rigor issue in Theorem 6.4. The proof counts minus-invariant edge subsets, then writes b_{m,n}=O(...) and immediately concludes the leading coefficient is non-integral. An upper bound alone does not determine a leading coefficient. The intended argument is probably fixable: for fixed n and large m, the number of non-matchings among those subsets is lower order, and the exact leading coefficient of the subset count is the sum S_n. But as written, the O-notation hides the needed equality. This is a minor patch compared to the missing CII/BI proofs.\n\nThe reference list looks standard and the paper does not lean on self-citations. Nothing circular jumps out. The author is upfront about the omitted proofs, which makes me think the gaps are likely fillable rather than signs of a wrong idea.\n\nWho is this for? People working on Borel orbits of symmetric varieties, quiver representations, or matching polynomials. The AIII section alone is worth reading. But a referee should send it back: the CII and BI parametrizations need complete proofs before the Can–Ugurlu conjecture can be accepted. I'd engage with the paper seriously, but I'd insist on the missing details before believing the full set of claims.","headline":"The type AIII matching bijection is genuine and well proved; the CII and BI sections are sketches and the Can–Ugurlu conjecture is not proven as written.","tokens_in":13683,"tokens_out":4259,"would_cite":true,"duration_ms":42809,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["05C70","20G05","14L30","05A20"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that Borel orbits in the symmetric varieties of types AIII and CII are naturally parametrized by matchings of corona graphs, and uses this to confirm the Can–Uğurlu conjecture on the non-integrality of an orbit-counting…","keywords":["Borel orbits","symmetric varieties","corona graph","double corona graph","matchings","quiver representations","ultra log-concavity","Can-Uğurlu conjecture"],"falsifier":"Compute the number of Borel orbits of $\\mathrm{SO}_{2m+2n+1}/S(\\mathrm{O}_{2m}\\times \\mathrm{O}_{2n+1})$ for a small case such as $m=n=1$ by direct linear algebra or by an existing clan classification, and compare it with the number of minus-invariant $2m$-matchings in $C_{2m+2n+1}$; a single mismatch would invalidate Theorem 5.16 and the conjecture proof.","tokens_in":12493,"feed_emoji":"🔗","tokens_out":7547,"duration_ms":57736,"temperature":0.7,"pith_summary":"The paper establishes that the Borel orbits in two families of classical symmetric varieties—type AIII and type CII—are naturally indexed by matchings in the corona graph $C_{m+n}$ and the double corona graph $C^{(2)}_{m+n}$. It uses this combinatorial description to prove that the orbit-counting sequences are symmetric, unimodal, and ultra log-concave, and to confirm a conjecture of Can and Uğurlu that the polynomial interpolating the number of Borel orbits in type BI has a non-integral leading coefficient for $n>0$. A sympathetic reader cares because the result turns complicated algebraic group orbit data into elementary graph-theoretic objects and yields concrete number-theoretic consequences.","feed_headline":"Borel orbits in symmetric varieties match corona-graph matchings","feed_subtitle":"The bijection yields ultra log-concavity and settles a conjecture on non-integral leading coefficients.","key_machinery":"The corona graph $C_n$ is obtained from the complete graph $K_n$ by attaching a new leaf $w_i$ to each vertex $v_i$; the double corona graph uses two parallel edges between each pair of vertices. The argument maps each Borel orbit to a binary-matrix representative, builds a complementary representation of the quiver $Q_p$ with dimension vector $d_{m,n}$, and applies Gabriel's theorem to decompose it into indecomposables corresponding to admissible roots. The minus involution on matchings, swapping $v_i \\leftrightarrow v_{-i}$ and $w_i \\leftrightarrow w_{-i}$, selects exactly the matchings that survive in the symplectic and orthogonal fixed-point situations.","core_discovery":"The central discovery is a bijection: for $(G,K) = (\\mathrm{GL}_{m+n}, \\mathrm{GL}_m \\times \\mathrm{GL}_n)$, each $B(k)$-orbit of $X = G/K$ corresponds to exactly one $m$-matching of the corona graph $C_{m+n}$; and for $(\\mathrm{Sp}_{2m+2n}, \\mathrm{Sp}_{2m} \\times \\mathrm{Sp}_{2n})$ with $\\mathrm{char}(k) \\neq 2$, each orbit corresponds to an $m$-matching of the double corona graph $C^{(2)}_{m+n}$. The correspondence passes through quiver representations of a Dynkin quiver of type $D_{m+n+2}$ with dimension vector $d_{m,n}$, which are translated into matchings of the underlying graph. In the orthogonal case (type BI), the same mechanism, applied to minus-invariant matchings, yields the count of Borel orbits of $\\mathrm{SO}_{2m+2n+1}/S(\\mathrm{O}_{2m}\\times \\mathrm{O}_{2n+1})$ and proves the Can–Uğurlu conjecture.","pith_inferences":["If the bijection is compatible with natural partial orders, the Bruhat order on these Borel orbit sets could be read directly from a poset of matchings under edge inclusion, giving a purely combinatorial handle on orbit closures.","The asymptotic method for extracting leading coefficients from matching counts is a general template: any symmetric variety whose orbits are counted by matchings in a dense graph will have an interpolating polynomial with a non-integral leading coefficient, provided the corresponding closed formula is not integer-valued.","The discrepancy between $k$-rational and $\\overline{k}$ orbits noted for type CII warns that analogous matching parametrizations for other symmetric varieties may require twisting by local systems when the field is not algebraically closed."],"forward_implications":["The orbit-counting sequences satisfy the recurrences $a_{p,m} = a_{p-1,m-1} + a_{p-1,m} + (p-1)a_{p-2,m-1}$ and $c_{p,m} = c_{p-1,m-1} + c_{p-1,m} + 2(p-1)c_{p-2,m-1}$, so the counts can be computed by elementary matching counts.","The matching description implies symmetric unimodality and ultra log-concavity of the orbit-counting sequences for types AIII and CII, via results on matching polynomials.","The type BI polynomial $b_{m,n}$ has degree $2n+1$ and non-integral leading coefficient for $n>0$, confirming the Can–Uğurlu conjecture.","The same matching mechanism is indicated to extend to other classical symmetric varieties arising from involutions of $\\mathrm{GL}_{m+n}$.","For type AIII the bijection holds over an arbitrary field $k$, not only over algebraically closed fields."],"supporting_citations":[{"why":"Supplies the Gabriel-type theorem used to decompose complementary quiver representations into admissible roots.","marker":"[4]"},{"why":"Established that $b_{m,n}$ is a polynomial in $m$ and stated the non-integrality conjecture that the paper proves.","marker":"[2]"},{"why":"Provides the Heilmann–Lieb matching polynomial facts used to derive unimodality and ultra log-concavity.","marker":"[3]"},{"why":"Introduced the recurrences for the orbit counts, which are recovered here from the matching description.","marker":"[1]"},{"why":"Showed the matching parametrization for type AI that motivated the extension to other symmetric varieties.","marker":"[7]"},{"why":"Gave the result that $\\phi$-fixed points form a single double coset, re-proved here through the Weyl-group lemma.","marker":"[11]"},{"why":"Provides the general theory of algebraic groups with involutions and the $\\tau$-normalizer description used in Section 5.1.","marker":"[9]"}],"fun_headline_variants":["Corona matchings count Borel orbits","Bijection: Borel orbits to corona matchings","Corona graph matchings settle symmetry conjecture","Ultra log-concavity via corona-orbit bijection","Orbits of symmetric varieties match corona graphs"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The parametrization of Borel orbits for the odd orthogonal symmetric variety (Theorem 5.16) is asserted by saying the symplectic arguments carry over with minimal modifications, and a key sufficiency step (Proposition 5.10) is stated without proof; the proof of the Can–Uğurlu conjecture rests directly on this unproved transfer.","fun_headline_variants_meta":{"raw":{"variants":["Corona matchings count Borel orbits","Bijection: Borel orbits to corona matchings","Corona graph matchings settle symmetry conjecture","Ultra log-concavity via corona-orbit bijection","Orbits of symmetric varieties match corona graphs"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000227,"raw_usage":{"total_tokens":1422,"prompt_tokens":849,"completion_tokens":573,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":465,"completion_tokens_details":{"reasoning_tokens":500}},"tokens_in":465,"tokens_out":573,"duration_ms":5389,"temperature":1.0,"reasoning_tokens":500,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T23:28:05.806342+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the number of Borel orbits of $\\mathrm{SO}_{2m+2n+1}/S(\\mathrm{O}_{2m}\\times \\mathrm{O}_{2n+1})$ for a small case such as $m=n=1$ by direct linear algebra or by an existing clan classification, and compare it with the number of minus-invariant $2m$-matchings in $C_{2m+2n+1}$; a single mismatch would invalidate Theorem 5.16 and the conjecture proof.","supporting_citations":[{"cited_title":"Lusztig, Canonical bases arising from quantized enveloping algebra s, Journal of the Amer- ican Mathematical Society 3 (1990), no","cited_arxiv_id":null,"evidence_quote":"Supplies the Gabriel-type theorem used to decompose complementary quiver representations into admissible roots."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Established that $b_{m,n}$ is a polynomial in $m$ and stated the non-integrality conjecture that the paper proves."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Heilmann–Lieb matching polynomial facts used to derive unimodality and ultra log-concavity."},{"cited_title":"The genesis of involutions (polarizations and lattice paths)","cited_arxiv_id":"1703.09881","evidence_quote":"Introduced the recurrences for the orbit counts, which are recovered here from the matching description."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Showed the matching parametrization for type AI that motivated the extension to other symmetric varieties."},{"cited_title":"Symmetric subgroup orbit closures on flag varieties: Their equivariant geometry, combinatorics, and connections with degeneracy loci","cited_arxiv_id":"1201.4397","evidence_quote":"Gave the result that $\\phi$-fixed points form a single double coset, re-proved here through the Weyl-group lemma."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the general theory of algebraic groups with involutions and the $\\tau$-normalizer description used in Section 5.1."}],"review_version":1}