{"id":"6b94a1c8-2167-4222-8fe6-64dda1c9b388","arxiv_id":"1908.09342","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The resonance graph of a plane elementary bipartite graph is constructible from an edge by peripheral convex expansions along a reducible face decomposition if and only if the infinite face is forcing.","lead":"This paper proves that the resonance graph of a plane elementary bipartite graph can be built from a single edge by repeatedly adding convex pieces exactly when the graph's infinite face is a forcing face. It supplies a clean criterion for which resonance graphs are peripheral convex expansions of an edge, generalizing earlier results.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 3.3's necessity direction leans on height formula (1) and Lemmas 2.4–2.5 from a benzenoid-paper [13], with no proof that these hold for all plane elementary bipartite graphs; if that transfer fails, the central characterization is unproved.","rationale":"Read in good faith: the theorem is plausible and the proof strategy is coherent. I attempted to break the internal steps. Lemma 3.2's 'choose x as the minimum element of (M(G;P−),≤)' looks suspicious, but it is repairable: the minimum a of the P−-sublattice must belong to M(G;P−,∂s), since a has no outgoing edges inside P− and yet must have a directed path down to the global minimum M̂0, which lies in P+; hence the first crossing edge of such a path is an edge of F incident with a. Similarly the 'It is clear' extension of M1,M2 to M'_1,M'_2 is valid by matching the even boundary cycle ∂G alternately, choosing the proper/improper matching according to the desired direction, with the interior cycles unchanged. Thus I do not regard those as load-bearing. The genuinely load-bearing assumption is the external transfer from [13]. Formula (1) and Lemmas 2.4 and 2.5 are used exactly in the necessity half of Theorem 3.3; no proof or generalization is supplied in this manuscript, and the cited paper's title indicates benzenoid systems. Since the class here is plane elementary bipartite graphs, which include non-benzenoid examples, this is a real scope gap. It is not an internal inconsistency, and it may be resolved by a positive test or by adding a lemma with a proof valid for the larger class. For that reason I keep the verdict conditional, same as the reader. If the concrete 4×4-grid check verifies all three identities, I would regard the gap as a presentational one and would support acceptance after a small revision; if it fails, the central theorem is false in the stated generality.","tokens_in":9328,"tokens_out":29086,"duration_ms":272092,"concrete_test":"Implement a brute-force check on the 4×4 grid (a plane elementary bipartite graph not covered by benzenoid-only results, if such a restriction exists): enumerate all perfect matchings, fix M̂0 and M̂1, and compute for every perfect matching M and every finite face f the value φ_M(f) = number of (M,M̂0)-alternating cycles in M⊕M̂0 containing f. Then verify (i) height(M(G)) = d(M̂1,M̂0) equals Σ_f φ_{M̂1}(f); (ii) for adjacent matchings M,M' with M⊕M' = ∂s, φ_M−φ_{M'} is 1 at s and 0 at all other faces; (iii) for a pair M'_1,M'_2 built as in Theorem 3.3's necessity proof, φ_{M'_1}(s)−φ_{M'_2}(s)=ψ_{M'_1M'_2}(s). Any failure is a concrete counterexample to the cited transfer; success would confirm the needed generalization.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3 derives formula (1), height(M(G)) = Σ_{f∈F} φ_{M̂1}(f), by invoking Theorem 3.2 of [13], and the contradiction in the necessity direction of Theorem 3.3 uses Lemma 2.4 (φ_{M'}(s)−φ_{M''}(s)=ψ_{M'M''}(s)) and Lemma 2.5 (cover iff φ changes by 1 on exactly one finite face) from the same paper. The cited paper is titled 'Resonance graphs and a binary coding for the 1-factors of benzenoid systems'; the present manuscript gives no argument that its statements, or their proofs, cover the strictly larger class of all plane elementary bipartite graphs. The graph class here includes non-hexagonal examples such as rectangular grids with interior vertices. If Lemmas 2.4–2.5 or the height formula are restricted to hexagonal systems, then the chain φ_{M̂1}(s)≥φ_{M'_1}(s)≥2 in the necessity proof has no basis, and the 'only if' direction of Theorem 3.3 collapses to an unproved assertion. The other two terse steps in the paper—the extension of interior perfect matchings to G and the 'choose x as the minimum' step in Lemma 3.2—can be justified from lattice/cut properties, so the scope of [13] is the load-bearing gap.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper characterizes, for a plane elementary bipartite graph G, when the resonance graph Z(G) can be obtained from a single edge by a sequence of peripheral convex expansions with respect to a reducible face decomposition. The main result (Theorem 3.3) states that this happens exactly when the infinite face of G is forcing. The proof works with the distributive lattice of perfect matchings of G, uses a height formula expressing the lattice height as a sum over finite faces of a function φ, and then relates equality in that height formula to the forcing-face condition and to the possibility of building Z(G) by peripheral convex expansions.","tokens_in":9664,"tokens_out":18014,"duration_ms":173540,"significance":"If the proof is made fully rigorous, the result is a clean and natural characterization that generalizes earlier work on outerplane bipartite graphs and catacondensed hexagonal systems. The connection between forcing faces and convex expansion structure is elegant, and the paper gives useful examples showing that the associated Θ-graph need not be a tree. The main proof strategy, reducing the problem to a height computation in the perfect-matching lattice, is well chosen. However, the correctness of the central implication depends on the applicability of several results from a cited paper whose title targets benzenoid systems, and on a few terse steps in the proof that need to be spelled out.","major_comments":[{"comment":"The proof relies on Theorem 3.2 and Lemmas 2.4 and 2.5 of [13] to establish the height formula (1) and the inequalities φ_{M'_1}(s) − φ_{M'_2}(s) = ψ_{M'_1M'_2}(s) and φ_{M̂1}(s) ≥ φ_{M'_1}(s). The cited paper is titled \"Resonance graphs and a binary coding for the 1-factors of benzenoid systems,\" and the present manuscript gives no argument that its statements, or their proofs, apply to the strictly larger class of all plane elementary bipartite graphs. Since this class includes non-hexagonal examples such as rectangular grids, the \"only if\" direction of Theorem 3.3 collapses if those results are in fact restricted to benzenoid systems. Please either quote the exact theorems from [13] confirming that they hold for all plane elementary bipartite graphs, or provide self-contained proofs of the needed statements.","section":"Section 3, Theorem 3.3, formula (1) and necessity direction"},{"comment":"The statement of Lemma 3.2 gives an equivalence with the equality height(M(G)) = height(M(H)) + 1, but the proof of Theorem 3.3 uses the stronger assertion that height(M(G_i)) ≥ height(M(G_{i-1})) + 1 for every reducible face decomposition, with equality exactly in the peripheral-expansion case. This inequality is not part of the lemma as stated, even though the proof of Lemma 3.2 suggests it. In addition, in the sufficiency proof of Lemma 3.2 the sentence \"In particular, we can choose x as the minimum element of the sublattice (M(G;P^−), ≤)\" is not justified as written: an arbitrary directed path's crossing edge need not have the minimum element as an endpoint, and one must show that a directed path can be chosen whose F-edge starts at that minimum and that the partner vertex in M(G;P^+,∂s) is not M̂0. Please state the inequality explicitly and supply the missing argument.","section":"Lemma 3.2 and sufficiency direction of Theorem 3.3"},{"comment":"The proof says \"It is clear that M1 (resp., M2) can be extended to a perfect matching M'_1 (resp., M'_2) of G such that both C and ∂G are proper M'_1-alternating and improper M'_2-alternating, and C is not contained in any (M'_1,M'_2)-alternating cycles other than ∂G.\" This extension step is load-bearing because it is used to produce a face s with ψ_{M'_1M'_2}(s) ≥ 2 and hence φ_{M'_1}(s) ≥ 2. The existence of such an extension is not immediate from the hypotheses, especially for an arbitrary perfect matching M1 of G − V(∂G), and the orientation compatibility with ∂G needs a proof. Please either prove this extension lemma or give a precise reference.","section":"Theorem 3.3, necessity direction, extension of M1 and M2"}],"minor_comments":[{"comment":"There is a typo: \"The we obtain a graph\" should read \"Then we obtain a graph.\"","section":"Introduction, first paragraph"},{"comment":"The abstract contains a spacing typo in \"elem entary\"; the same word is also broken across lines in the introduction. Please fix the formatting.","section":"Abstract and Introduction"},{"comment":"The notation M(G;P^-, \\overline{\\partial s}) is introduced without defining the overline notation. Please define it explicitly as the complement of the set M(G;P^-,∂s).","section":"Section 2, notation for M(G;P^−,∂s)"},{"comment":"The sentence \"By Theorem 3.2 in [13] states that ...\" is grammatically awkward and should be rewritten as \"By Theorem 3.2 of [13], ... .\"","section":"Proof of Theorem 3.3"},{"comment":"Since the proof of Theorem 3.3 uses the strict inequality height(M(G_i)) > height(M(G_{i-1})) + 1 in the non-peripheral case, it would be clearer to incorporate this inequality into the statement of Lemma 3.2 or to state it as a separate corollary.","section":"Lemma 3.2"}],"recommendation":"major_revision","confidential_remarks":"The central theorem is plausible and the proof strategy is sound, but the manuscript's heavy reliance on [13] is a genuine risk: the editor may want an additional check from a referee familiar with the exact contents of [13] to confirm that its Theorem 3.2 and Lemmas 2.4 and 2.5 are proved for all plane elementary bipartite graphs and not only for benzenoid systems. If that confirmation is obtained, the other two main gaps appear fixable with added arguments and a strengthened Lemma 3.2. The paper would then be a solid contribution to the resonance-graph literature."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: Theorem 3.3 is a real advance over the outerplane and catacondensed results, and the equivalence with \"the infinite face is forcing\" is a clean, memorable criterion. The proof route — height formula plus the cover lemma — is coherent, and the sufficiency argument is elegant once Lemma 3.2 is granted. I think the paper deserves a serious referee, but the referee should press hard on one thing: the unqualified use of [13].\n\nThe novelty is genuine. The characterization is not contained in the earlier work, and the class of plane elementary bipartite graphs is substantially wider than benzenoid systems. The author is careful to distinguish the new result from [3] and [11] in the final section.\n\nThe main soft spot is the citation transfer. Formula (1) and Lemmas 2.4 and 2.5 are taken from a paper whose title says \"benzenoid systems.\" The present manuscript never argues that those results, or their proofs, hold for arbitrary plane elementary bipartite graphs. The necessity direction of Theorem 3.3 leans on this: the chain φ_{M̂1}(s) ≥ φ_{M'1}(s) ≥ 2 is built on the difference lemma and the cover criterion. If those lemmas are in fact restricted to hexagonal systems, the only-if direction does not go through. This is not a demonstrated error; it is an unstated assumption. A referee can ask for a proof or a broader citation, and it may well be that the lemmas generalize without new ideas. But the burden is on the author.\n\nThe other terse steps are less worrying. The unique extension of a perfect matching of H to one of G in M(G;P−) is a short parity/path argument. The \"choose x as the minimum\" in Lemma 3.2 is a bit hand-wavy but can be justified by taking a path that goes through the minimum of the sublattice before crossing the face-label edge. Those are presentation gaps.\n\nBottom line: this is a paper for the resonance graph and mathematical chemistry crowd, not a broad-audience result. It deserves peer review. I'd recommend asking the author to clarify the scope of the cited results and clean up Lemma 3.2. My own verdict would hinge on the answer to the [13] question.","headline":"A clean characterization of resonance graphs that genuinely generalizes known results, but the proof carries a citation-scope gap that a referee should pin down before the theorem is trusted.","tokens_in":10119,"tokens_out":13284,"would_cite":true,"duration_ms":124570,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["05C70","05C10","05C75"],"pacs":[],"model":"deepseek-v4-flash","headline":"Resonance graphs of plane elementary bipartite graphs grow from a single edge by peripheral convex expansions exactly when the infinite face is forcing.","keywords":["resonance graph","peripheral convex expansion","reducible face decomposition","forcing face","plane elementary bipartite graph","distributive lattice","median graph","Z-transformation graph"],"falsifier":"Compute a plane elementary bipartite graph with a non-hexagonal finite face and test whether $d_{Z(G)}(\\hat M_1,\\hat M_0) = \\sum_{f\\in F} \\varphi_{\\hat M_1}(f)$; any inequality would falsify the imported height formula and therefore the necessity proof. More directly, find a graph whose infinite face is forcing but whose resonance graph cannot be assembled from an edge by peripheral convex expansions along any reducible face decomposition, which would disprove Theorem 3.3.","tokens_in":9096,"feed_emoji":"🧩","tokens_out":12823,"duration_ms":100544,"temperature":0.7,"pith_summary":"The paper proves a characterization of when a resonance graph can be assembled from a single edge. For a plane elementary bipartite graph $G$, the resonance graph $Z(G)$ — whose vertices are perfect matchings of $G$ and whose edges join two matchings that differ exactly around one finite face — can be obtained from an edge by a sequence of peripheral convex expansions following a reducible face decomposition of $G$ if and only if the infinite face of $G$ is forcing. A forcing infinite face means its boundary is an even cycle whose removal leaves at most one perfect matching. The result matters because it reduces a global structural question about a potentially large graph of matchings to a local, checkable condition on one face, and it gives an explicit construction of the whole resonance graph from that condition.","feed_headline":"Resonance graphs grow from one edge iff outer face is forcing","feed_subtitle":"The perfect-matching graph has a simple edge-by-edge construction when the outer face fixes all matchings.","key_machinery":"The load-bearing object is the resonance graph $Z(G)$ and the expansion operation that builds it. A peripheral convex expansion is a convex expansion in which one of the two isometric subgraphs is the whole starting graph and the other is a convex subgraph; the new graph is formed by taking disjoint copies of the two subgraphs and adding an edge between corresponding vertices of their intersection. A reducible face decomposition builds $G$ face by face, each step adding one finite face by an odd-length boundary path. The proof's central mechanism is Lemma 3.2, which converts the expansion into a one-step lattice-height statement: $Z(G)$ is a peripheral convex expansion of $Z(H)$ if and only if $\\operatorname{height}(M(G)) = \\operatorname{height}(M(H)) + 1$. The height formula (1), summing the face contributions $\\varphi_{\\hat M_1}(f)$, ties this lattice height back to the geometry of alternating cycles and forces the all-contributions-equal-one condition.","core_discovery":"The central claim is Theorem 3.3: $Z(G)$ is constructible from an edge by peripheral convex expansions with respect to a reducible face decomposition of $G$ exactly when the infinite face of $G$ is forcing. The proof works through the distributive lattice of perfect matchings. It uses the height formula $\\operatorname{height}(M(G)) = \\sum_{f\\in F} \\varphi_{\\hat M_1}(f)$, where $\\varphi_{\\hat M_1}(f)$ counts the $(\\hat M_1,\\hat M_0)$-alternating cycles whose interior contains the finite face $f$, and shows via Lemma 3.2 that each peripheral convex expansion over a reducible face raises the lattice height by exactly one. Hence a graph built from an edge in $n$ steps has height $n$, and the forcing condition is exactly what makes every finite face contribute one to the height; a non-forcing infinite face forces some face to contribute at least two, which blocks the construction.","pith_inferences":["Inference: The imported height formula (1) is the true bottleneck of the proof; verifying it on a non-benzenoid plane elementary bipartite graph would show whether the characterization extends, needs a different proof, or fails outside the hexagonal setting.","Inference: Because forcing of the infinite face is equivalent to $\\hat M_0 \\oplus \\hat M_1 = \\partial G$, the theorem yields a polynomial-time certificate for constructibility: check that the symmetric difference of the two extremal perfect matchings is exactly the outer boundary.","Inference: Since resonance graphs are median graphs, the expansion description suggests that constructible resonance graphs are exactly those median graphs whose $\\Theta$-classes can be ordered level by level to match a reducible face decomposition; this is a testable structural conjecture, not a claim of the paper."],"forward_implications":["If the infinite face of $G$ is forcing, Theorem 3.3 supplies an explicit inductive construction of the entire resonance graph $Z(G)$ starting from a single edge.","For a plane elementary bipartite graph with $n$ finite faces, constructibility is equivalent to $\\operatorname{height}(M(G)) = n$ and to $\\varphi_{\\hat M_1}(f) = 1$ for every finite face $f$.","The paper's Corollary 3.4 says that if every finite face of $G$ has a vertex on the outer boundary, then $Z(G)$ is constructible from an edge.","The constructibility condition does not by itself force the graph $\\Theta(Z(G))$ to be isomorphic to the inner dual of $G$; the paper gives two examples and leaves that stronger characterization as Question 3.1."],"supporting_citations":[{"why":"Supplies the height formula (1), the face-contribution function $\\varphi_M(f)$, and Lemmas 2.4 and 2.5 used in the necessity direction of Theorem 3.3.","marker":"[13]"},{"why":"Provides Theorem 2.2, the convex expansion structure of $Z(G)$ over a reducible face, including the $\\Theta$-class and the peripheral-expansion criterion.","marker":"[2]"},{"why":"Gives the reducible-face machinery: the common boundary of a reducible face and $G$ is an odd path, and the outer boundary is alternating with respect to the extremal matchings.","marker":"[10]"},{"why":"Establishes that a plane bipartite graph with more than two vertices is elementary if and only if it has a reducible face decomposition, which anchors the construction sequence.","marker":"[15]"},{"why":"Shows the perfect matchings form a finite distributive lattice whose Hasse diagram is the resonance digraph, used for the height and Jordan-Dedekind arguments.","marker":"[9]"},{"why":"Defines forcing faces, the property whose status for the infinite face the theorem ties to constructibility.","marker":"[4]"}],"fun_headline_variants":["Resonance graphs from one edge iff outer face is forcing","Forcing outer face enables edge-by-edge resonance graph construction","Edge-by-edge growth: resonance graphs require forcing outer face","Single-edge build for resonance graphs when outer face is forcing"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The necessity proof assumes that the height formula and Lemmas 2.4 and 2.5 from reference [13], originally proved for benzenoid systems, hold for every plane elementary bipartite graph; if those results are limited to hexagonal systems, the contradiction argument that every finite face contributes exactly one to the height collapses.","fun_headline_variants_meta":{"raw":{"variants":["Resonance graphs from one edge iff outer face is forcing","Forcing outer face enables edge-by-edge resonance graph construction","Edge-by-edge growth: resonance graphs require forcing outer face","Single-edge build for resonance graphs when outer face is forcing"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000904,"raw_usage":{"total_tokens":3803,"prompt_tokens":776,"completion_tokens":3027,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":392,"completion_tokens_details":{"reasoning_tokens":2959}},"tokens_in":392,"tokens_out":3027,"duration_ms":24337,"temperature":1.0,"reasoning_tokens":2959,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:15:12.589972+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute a plane elementary bipartite graph with a non-hexagonal finite face and test whether $d_{Z(G)}(\\hat M_1,\\hat M_0) = \\sum_{f\\in F} \\varphi_{\\hat M_1}(f)$; any inequality would falsify the imported height formula and therefore the necessity proof. More directly, find a graph whose infinite face is forcing but whose resonance graph cannot be assembled from an edge by peripheral convex expansions along any reducible face decomposition, which would disprove Theorem 3.3.","supporting_citations":[{"cited_title":"Zhang, P","cited_arxiv_id":null,"evidence_quote":"Supplies the height formula (1), the face-contribution function $\\varphi_M(f)$, and Lemmas 2.4 and 2.5 used in the necessity direction of Theorem 3.3."},{"cited_title":"Che, Structural properties of resonance graphs of plane e lementary bipartite graphs, Discrete Appl","cited_arxiv_id":null,"evidence_quote":"Provides Theorem 2.2, the convex expansion structure of $Z(G)$ over a reducible face, including the $\\Theta$-class and the peripheral-expansion criterion."},{"cited_title":"Taranenko and A","cited_arxiv_id":null,"evidence_quote":"Gives the reducible-face machinery: the common boundary of a reducible face and $G$ is an odd path, and the outer boundary is alternating with respect to the extremal matchings."},{"cited_title":"Zhang and F","cited_arxiv_id":null,"evidence_quote":"Establishes that a plane bipartite graph with more than two vertices is elementary if and only if it has a reducible face decomposition, which anchors the construction sequence."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows the perfect matchings form a finite distributive lattice whose Hasse diagram is the resonance digraph, used for the height and Jordan-Dedekind arguments."},{"cited_title":"Che and Z","cited_arxiv_id":null,"evidence_quote":"Defines forcing faces, the property whose status for the infinite face the theorem ties to constructibility."}],"review_version":1}