{"id":"22cee2a5-f863-411e-880c-06cc10b020a2","arxiv_id":"2506.16333","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A tropical linear series is combinatorial limit if and only if it is structured, and every strongly recursive tropical linear series is combinatorial limit; the reverse inclusion fails from rank three onward.","lead":"This paper compares two definitions of linear series on tropical curves and proves that every strongly recursive tropical linear series is a combinatorial limit linear series. It also gives a simpler characterization of the second class and shows where the two definitions diverge for rank at least three.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Definition 2.14 and Corollary 3.13 disagree: local arrays are redundant closures, not permutation arrays, so the r-slope structure asserted for Theorem 1.1 need not exist as defined.","rationale":"I agree with the reader's conditional assessment. The main theorem's proof depends on Corollary 3.13, and its assertion that the local array is a permutation array contradicts Definition 2.14 and the paper's own Theorem 3.11. The rank-1 example makes the mismatch concrete, so this is not a judgment call about a subtle variant. I nevertheless do not recommend rejection: the construction is repairable by using the redundant closure or rank array as the slope datum, and the surrounding arguments provide enough scaffolding that the intended theorem is plausible. The verdict should remain conditional pending either a rewrite of Definition 2.14 or of Corollary 3.13, plus a proof of Theorem 5.2, which is currently asserted without proof. I did not find a separate fatal flaw; the other concerns I considered, such as details in the limiting argument of Theorem 4.4, appear less central and likely fixable.","tokens_in":21338,"tokens_out":11196,"duration_ms":110504,"concrete_test":"Check the disputed step on the rank-1 interval series of Example 2.4. Write down Definition 2.14(3) and the local arrays from Example 3.1; at an interior point the array is {(0,1),(1,0),(0,0)}. Under the EL00 definition used in Section 2.3, (0,0) is a redundant point, so this array is not a permutation array; if one instead takes P_v = {(0,1),(1,0)}, the function with local slopes (0,0) is not compatible, contradicting Σ ⊆ R(D,S). This single example settles whether Corollary 3.13's assertion that P_v is a permutation array holds with the paper's definitions. A secondary check would be to re-derive the statement of Theorem 5.2 from a corrected Definition 2.14 and the proof of Theorem 4.4, since the paper only says that the results extend without giving the proof.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Definition 2.14(3) requires each P_v to be a permutation array in the EL00 sense adopted in Section 2.3, i.e. a totally rankable dot array with no redundant points dotted. Theorem 3.11 proves that the local array of a strongly recursive tropical linear series is the redundant closure of a permutation array, and the examples in Figures 4 and 5 show such arrays can contain redundant points. Corollary 3.13 then sets P_v equal to this local array and calls it a permutation array; under Definition 2.14 this is false. If P_v is instead taken to be the minimal permutation array inside the local array, the compatibility condition ∂_v(f) ∈ P_v in Definition 2.14 excludes functions at redundant positions, so Σ ⊆ R(D,S) fails. The rank-1 interval series of Example 2.4 already exhibits this: at interior valence-2 points the local array contains the redundant point (0,0), so the strict permutation array does not contain all local data of Σ. Thus the proof of Theorem 1.1 currently has no valid r-slope structure as defined. The defect is repairable by defining the slope datum as the rank array or redundant closure, but as written Definition 2.14 and Corollary 3.13 are inconsistent. The same definition is used in the structured-series statement of Section 5.1, so this concern is load-bearing for the paper's central inclusion, not a peripheral typo.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper compares several recent definitions of linear series on tropical curves: strongly recursive tropical linear series (Farkas–Jensen–Payne), tropical linear series in the sense of Chang et al., and combinatorial limit linear series (Amini–Gierczak). The main result, Theorem 1.1, asserts that every strongly recursive tropical linear series of rank r is a combinatorial limit linear series of rank r. The proof proceeds by studying the local arrays of strongly recursive series, showing these arrays are redundant closures of permutation arrays, and then constructing a slope structure and verifying the Baker–Norine and local rank properties. Section 5 extends the main theorem to structured tropical linear series, yielding Theorem 5.2, an equivalence between structured tropical linear series and combinatorial limit linear series. The paper also gives counterexamples to the converse of Theorem 1.1 for rank at least 3, discusses when permutation arrays are realizable as local arrays, and proves low-rank and low-dimensional equivalence results.","tokens_in":21571,"tokens_out":18318,"duration_ms":182026,"significance":"If Theorems 1.1 and 5.2 are correct, the paper achieves a genuine unification of two major recent frameworks for tropical linear series: the recursive definition of Farkas–Jensen–Payne and the slope-structure definition of Amini–Gierczak. The paper is careful and well illustrated, with useful examples (Figures 5 and 6), a clear account of permutation arrays, and explicit low-rank results. It also provides concrete counterexamples, drawn from the Vámos matroid and from Billey–Vakil, showing that certain natural converse statements fail. The machine-checked combinatorial assertions in Section 6 are a further strength. However, the current proof of Theorem 1.1 contains a definitional mismatch between the slope structure in Definition 2.14 and the local arrays produced in Corollary 3.13, and this mismatch is load-bearing for the main theorem.","major_comments":[{"comment":"The r-slope structure S in Corollary 3.13 is not well-defined as written. Definition 2.14(3) requires each P_v to be a permutation array in the strict sense of Subsection 2.3, that is, a totally rankable array with no redundant points dotted. Theorem 3.11, however, proves only that the local array of a strongly recursive tropical linear series is the redundant closure of a permutation array, and Example 3.1 and Figure 5 show that this closure can contain redundant points, e.g. the point (0,0) at the valence-2 points in Example 2.4. Corollary 3.13 then sets P_v equal to the local array and asserts that it is a permutation array; under Definition 2.14 this is false. If one instead takes P_v to be the minimal permutation array inside the local array, the compatibility condition ∂_v(f) ∈ P_v excludes functions that realize redundant local data, so the inclusion Σ ⊆ R(D,S) fails. Thus the proof of Theorem 1.1 has no valid r-slope structure as currently defined. This is repairable by changing the slope-structure definition to use rank arrays or redundant closures rather than strict permutation arrays, but as written Definition 2.14 and Corollary 3.13 are inconsistent.","section":"Section 3, Corollary 3.13; Definition 2.14"},{"comment":"The advertised equivalence is asserted without a proof. The sentence \"The results of Section 4 and the proof of Theorem 1.1 extend also to structured tropical linear series\" is not a proof of Theorem 5.2. The nontrivial direction, that every structured tropical linear series is a combinatorial limit linear series, requires verifying the admissibility condition of Definition 2.16, in particular the existence, for every effective divisor E of degree r, of a function satisfying both (BNRP) and (LRP). While the induction in Theorem 4.4 likely adapts, the base case uses condition (1) of Definition 2.2 and the induction step uses the slope structure, and these points should be checked explicitly. As written, the main characterization rests on an unproved assertion.","section":"Section 5.1, Theorem 5.2"}],"minor_comments":[{"comment":"The last sentence of the abstract is duplicated: \"Finally, address the realizability of permutation arrays as local arrays of linear series on tropical curves\" appears twice.","section":"Abstract"},{"comment":"Remark 5.1 says that a combinatorial limit linear series is by definition a structured tropical linear series. This is true only after checking that conditions (1) and (2) of Definition 2.2 hold for an admissible submodule; the implication should be stated explicitly rather than left to the reader.","section":"Section 5.1, after Theorem 5.2"},{"comment":"The sentence \"Consider the rank rd tropical linear series R(rd·v)\" appears to contain a typo or a gap: R(rd·v) has rank rd, not r, and the assertion that any r+1 of its elements are contained in a strongly recursive tropical linear subseries of rank r does not follow directly from condition (3) of Definition 2.3, which applies to sets of size equal to the rank. Please clarify the intended statement and argument.","section":"Section 6.1, Proposition 6.1"},{"comment":"The compactness argument extracts a convergent subsequence but does not explicitly justify that the needed limits of slopes at every tangent vector exist. Since Lemma 4.2 requires the existence of such limits, a brief diagonal argument over the finite slope data, or an explicit statement that the slopes are bounded integers and stabilize on a further subsequence, would make the proof fully rigorous.","section":"Section 4, proof of Theorem 4.4"}],"recommendation":"major_revision","confidential_remarks":"The central idea of the paper is sound and the comparison of definitions is valuable, but the mismatch between Definition 2.14 and Corollary 3.13 must be resolved before the main theorem can be accepted. I would recommend that the author consult the original formulation in Amini–Gierczak to determine whether the slope structure should be stated in terms of rank arrays rather than strict permutation arrays, and then revise Corollary 3.13 and Theorem 5.2 accordingly. No concerns about novelty or attribution; the paper is a genuine contribution to the area."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read this one if you work on tropical linear series. Burkholder proves that strongly recursive tropical linear series are combinatorial limit linear series (Theorem 1.1), gives a cleaner characterization via slope structures, and constructs counterexamples to the converse for rank at least 3 using Vámos matroids. That is new content, not a survey, and the main line of argument is coherent. The local-array technology in Section 3 is the right tool, and Theorem 3.11—local arrays are redundant closures of permutation arrays—is the key insight.\n\nWhere the paper gets soft:\n\n1) Theorem 5.2, the advertised equivalence 'structured iff combinatorial limit linear series,' is asserted after a two-sentence remark that the results 'extend.' That is not a proof. The only-if direction is essentially definitional, but the if direction is exactly the substance of Theorem 1.1 plus a weakened recursion assumption, and the reader is left to fill in an induction that is not written. This needs a real proof or an explicit reduction.\n\n2) Definition 2.14 and Corollary 3.13 do not agree. Definition 2.14(3) asks for a permutation array at each vertex in the strict EL00 sense, no redundant points dotted. Theorem 3.11 shows local arrays are redundant closures, which by definition include redundant points (Example 2.4 interior points already show (0,0) in the local array). Corollary 3.13 calls the local array itself a permutation array. If you take the strict permutation array inside it, Σ⊆R(D,S) fails because functions at redundant dots are excluded. If you loosen the definition to allow redundant points, the slope structure in Corollary 3.13 works and Theorem 1.1 survives, but as written the paper is internally inconsistent. This is repairable—define slope structures via rank arrays or redundant closures—but it is load-bearing for the proof of Theorem 1.1, not cosmetic.\n\nMinor points: the computer checks in Section 6 are reported without code or certificates; for finite checks this is a minor reproducibility issue. Some dependence on [CDI+] and [Wan24] is heavier than the text admits; those are the right citations, but the reader needs them at hand.\n\nBottom line: the central inclusion is likely true and the framework is a genuine advance; the paper needs a fix to the slope-structure definition and a real proof of Theorem 5.2 before it is accepted. I would send it to a serious referee—ideally one who works on matroids or tropical linear series—rather than desk-reject. It is well above threshold for refereeing, but not ready as is.","headline":"A genuinely useful unification of tropical linear series definitions, with one definitional mismatch and one missing proof that are fixable; worth serious refereeing.","tokens_in":22089,"tokens_out":3306,"would_cite":true,"duration_ms":33447,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14T10","14H51"],"pacs":[],"model":"deepseek-v4-flash","headline":"Structured tropical linear series equal combinatorial limit series","keywords":["tropical linear series","strongly recursive tropical linear series","combinatorial limit linear series","permutation arrays","slope structures","metric graphs","tropical submodules","rank arrays"],"falsifier":"Inspect the paper's own Figure 5 example at an interior point of an interval: the local array has three dotted points in a $2\\times 2$ square, with the corner $(0,0)$ equal to the coordinatewise minimum of the other two. If that corner counts as an extra filled-in point and the slope-structure definition requires the stricter array with no such points, then Corollary 3.13's constructed structure is not a valid slope structure, and Theorem 1.1 is not established by the given proof.","tokens_in":21090,"feed_emoji":"🔗","tokens_out":10912,"duration_ms":97718,"temperature":0.7,"pith_summary":"This paper attempts to settle whether two recently proposed definitions of linear series on tropical curves describe the same objects. It proves that every strongly recursive tropical linear series of rank $r$ is a combinatorial limit linear series of rank $r$, and then proves the sharper equivalence: a tropical linear series is structured (all of its functions fit inside a fixed slope structure) if and only if it is a combinatorial limit linear series. That matters because the two definitions encode different intuitions, one built from tropical vector-space-like closure and recursive containment, the other from local slope and permutation data at every point; matching them lets results proved in either language carry over. The paper also shows the converse of its main theorem fails for strongly recursive series in ranks $r\\ge 3$, so the coincidence is not unconditional: the structured condition is exactly the extra hypothesis that closes the gap.","feed_headline":"Structured tropical linear series equal combinatorial limit series","feed_subtitle":"The two definitions of tropical linear series coincide once a structured condition is added.","key_machinery":"The load-bearing object is the local array of a tropical linear series at a point $v$ of the metric graph: the subset of $[r]^d$ recording, for every tangent direction at $v$, which of the $r+1$ slope values is attained by some function in the series. Four closure properties, closure under coordinatewise minima (P1), realization of every slope at every tangent vector (P2), tropical dependence of any $r+2$ elements (P3), and the recursive containment condition (P4), force the array to be the redundant closure of a unique rank-$r$, dimension-$d$ permutation array; here a permutation array is a minimal totally rankable dot array with no redundant dotted positions, and its redundant closure adds all points obtainable as coordinatewise minima of its entries. That description is what lets the paper attach to the graph an $r$-slope structure whose pointwise data are permutation arrays, turning the local data into the admissible-submodule form required by the combinatorial limit definition. The limit argument of Section 4 then supplies the Baker-Norine and local rank properties for effective divisors of degree $r$, completing the bridge between the two definitions.","core_discovery":"The central claim is that the recursive definition of a tropical linear series, augmented by the structured condition that the whole series lies inside a slope structure, is identical to the combinatorial limit definition. Theorem 1.1 proves that any rank-$r$ strongly recursive tropical linear series is a combinatorial limit linear series of rank $r$, and Theorem 5.2 upgrades this to an equivalence: structured tropical linear series are exactly combinatorial limit linear series, so the local-rank condition in the original combinatorial definition can be dropped. The proof route is local: at each point of the metric graph, the series determines a local array of slope-index vectors, and the paper proves this array is the redundant closure of a unique permutation array (Theorem 3.11), which yields the slope structure needed for admissibility. A separate limit argument (Theorem 4.4) produces, for every effective divisor $E$ of degree $r$, a function satisfying both the Baker-Norine rank property and the local rank property for $E$. The converse of Theorem 1.1 fails for ranks $r\\ge 3$, and the paper isolates condition (3) of the recursive definition as the obstruction.","pith_inferences":["Beyond the paper: if every tropical linear series turns out to be structured, then the two definitions coincide on all tropical linear series; the paper's own open Question 5.3 is the precise test of this.","Beyond the paper: the rank-3 counterexample suggests a general matroid-theoretic obstruction to strong recursion; one could check whether the Levi intersection property characterizes exactly when the recursive condition (3) survives.","Beyond the paper: the rank-2 realizability checks raise the possibility that all rank-2 combinatorial limit linear series are strongly recursive, which would resolve Question 5.11 with a constructive proof for all dimensions."],"forward_implications":["Every strongly recursive tropical linear series of rank $r$ automatically satisfies the admissibility conditions of a combinatorial limit linear series of rank $r$.","A tropical linear series is structured exactly when it is a combinatorial limit linear series, giving a shorter definition that drops the separate local-rank requirement.","On the interval and loop metric graphs, and in ranks $0,1,2$ on any metric graph, the tropical-linear-series and combinatorial-limit definitions coincide.","The converse direction fails in ranks $r\\ge 3$: there exist combinatorial limit linear series that are not strongly recursive.","Some permutation arrays, including a rank-$3$ array in dimension $4$, cannot be local arrays of strongly recursive tropical linear series, so the local-array description does not extend to all permutation arrays."],"supporting_citations":[{"why":"defines combinatorial limit linear series, slope structures, and admissible submodules, and supplies the topological-closure fact used in the final step of Theorem 1.1","marker":"[AG22]"},{"why":"provides the theory of totally rankable dot arrays, permutation arrays, redundant closures, and the characterization used to prove Theorem 3.11","marker":"[EL00]"},{"why":"introduces strongly recursive tropical linear series and the lemma fixing the ordered set of slopes along each tangent vector","marker":"[FJP25]"},{"why":"supplies the tropical-dependence lemma and the finite pivot lemma used to control generic local arrays","marker":"[JP22]"},{"why":"gives the construction of tropical linear series from valuated matroids and the matroid examples used to build counterexamples to the converse for ranks at least 3","marker":"[CDI+]"},{"why":"provides the free-extension argument that lifts the rank-3 counterexamples to all higher ranks","marker":"[Wan24]"},{"why":"supplies the rank-3 dimension-4 permutation array used to show that some permutation arrays cannot be local arrays of strongly recursive tropical linear series","marker":"[BV08]"}],"fun_headline_variants":["Structured tropical series equal combinatorial limits","Structured recursive series match combinatorial limits","Two linear series definitions unify on tropical curves","Structured condition unifies tropical linear series definitions"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof assumes that filling in all the extra points obtained by combining entries of a minimal local pattern still counts as the pattern required by the slope structure; if the stricter reading of the definition is used, the constructed slope structure may be invalid and the main theorem is not proved as written.","fun_headline_variants_meta":{"raw":{"variants":["Structured tropical series equal combinatorial limits","Structured recursive series match combinatorial limits","Two linear series definitions unify on tropical curves","Structured condition unifies tropical linear series definitions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00082,"raw_usage":{"total_tokens":3559,"prompt_tokens":884,"completion_tokens":2675,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":500,"completion_tokens_details":{"reasoning_tokens":2620}},"tokens_in":500,"tokens_out":2675,"duration_ms":19039,"temperature":1.0,"reasoning_tokens":2620,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T19:29:47.298687+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Inspect the paper's own Figure 5 example at an interior point of an interval: the local array has three dotted points in a $2\\times 2$ square, with the corner $(0,0)$ equal to the coordinatewise minimum of the other two. If that corner counts as an extra filled-in point and the slope-structure definition requires the stricter array with no such points, then Corollary 3.13's constructed structure is not a valid slope structure, and Theorem 1.1 is not established by the given proof.","supporting_citations":[],"review_version":2}