{"id":"ceea2a0c-02fa-4f65-a463-d2f369ac36fa","arxiv_id":"2511.18545","paper_version":4,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For each genus g, the set of all separating semigroups of real curves of genus g is finite.","lead":"This paper proves that for each fixed genus, only finitely many different separating semigroups can occur for real algebraic curves of that genus. It gives a structural decomposition of the semigroup into finite special data plus finitely many cones, and bounds the degrees of the cone generators.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Unproved strengthening of Orevkov's Lemma 3.2 (converse for special divisors) is load-bearing for Theorem 1 and the degree bound in Theorem 2.","rationale":"The reader's weakest_assumption is exactly the gap I find most load-bearing. The footnote following Lemma 1 is the only place in the paper where a nontrivial strengthening is asserted without proof. The rest of the argument — the linear algebra in Theorem 1, Dickson's lemma in Lemma 4, the counting in Lemma 4, and the uniform bound via Lemma 6 — is sound modulo this point. I checked whether the strengthened converse is actually unnecessary for Lemma 6: one could try to apply a restricted version of Theorem 1 to non-special P_i and only conclude Q_i is non-special, using the original non-special converse. However, the paper does not state or prove such a restricted theorem; it relies on the full Lemma 1 as stated. Moreover, Corollary 1 and Theorem 1 as publicized genuinely require the strengthened converse. Thus the gap is not merely cosmetic. I do not see a separate flaw in the proof of Theorem 1's linear algebra: the existence of a relation with mixed signs and a positive combination with at least half the indices is plausible via a dimension argument, and rescaling tangent vectors preserves positivity. The overlap issue in Lemma 4's disjoint decomposition is a minor technical slip because the final union still covers Sep(X), and Question 1 acknowledges the potential non-disjointness. Therefore my concern aligns with the reader's, and the conditional verdict should stand.","tokens_in":7756,"tokens_out":27320,"duration_ms":275665,"concrete_test":"Obtain the proof of [Ore19, Lemma 3.2] and identify exactly where h^1(P)=0 is used. Then, on the genus 3 non-hyperelliptic curve of §3.3, compute a basis of holomorphic differentials and, for all effective divisors of degree 3 and 4 with distinct real points, test whether there exist positive tangent vectors satisfying (†). Compare each such divisor's degree partition with the known Sep(X) from §3.3. If any divisor satisfies (†) but is not separating, the strengthened converse is false and Theorem 1 collapses; if all such divisors are separating, the removal is supported for the relevant special cases.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Lemma 1 states an iff characterization of separating divisors using positive tangent vectors satisfying (†). The (⇐) direction is attributed to Orevkov's Lemma 3.2, but a footnote admits the original statement assumed the divisor is non-special and asserts this assumption 'can be removed' without proof. This strengthened converse is used in the proof of Theorem 1: after deleting points, the subdivisor Q is shown to satisfy (†), and Lemma 1(⇐) makes Q separating. Q may be special — for minimal separating divisors of degree g+1, the subdivisors can have degree near half and need not be non-special. If the removal is false, Theorem 1 fails as stated; Corollary 1 (minimal separating degree ≤ g+1) and Lemma 6's bound deg P_i ≤ 4g−3, the key to Theorem 2's finiteness, both rest on it. The paper gives no proof of the strengthening, only a footnote. This is the single most load-bearing unproved assertion in the argument.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies the separating semigroup Sep(X) of a real algebraic curve X, consisting of the degree vectors of separating morphisms to P^1. The main result (Theorem 2) is that for every genus g there are only finitely many possible separating semigroups of genus-g curves. The proof scheme is as follows: Lemma 4 expresses Sep(X) as the union of a finite set Sep_s(X) of degree partitions of special separating divisors and finitely many cones d(P_i)+N_0^r attached to minimal non-special separating divisors; Theorem 1 is a deletion result showing that every separating divisor of degree n≥g+2 contains a proper separating subdivisor of degree at least ⌈n/2⌉; and Lemma 6 uses Theorem 1 to prove the uniform bound deg P_i≤4g−3 for the minimal non-special divisors, which yields finiteness. The paper also proves that Sep(X) is never finitely generated when b0(RX)≥2 and illustrates the decomposition on examples.","tokens_in":8051,"tokens_out":23586,"duration_ms":215364,"significance":"If the proof is completed, the main result is a valuable finiteness theorem: it reduces the classification of separating semigroups to finite data and gives a uniform explanation for the known case-by-case results. The linear-algebra deletion argument is elegant and potentially applicable to other questions about real divisors. The paper is clearly written and the structural decomposition is a useful conceptual contribution. However, the central argument currently rests on an unproved strengthening of Orevkov's Lemma 3.2 (the converse for special divisors), so the claims are not fully established in the present form.","major_comments":[{"comment":"The equivalence in Lemma 1 is stated without qualification, but footnote 1 concedes that the (⇐) direction of Orevkov's Lemma 3.2 was proved only for non-special divisors, and asserts the assumption 'can be removed' without proof or reference. This is load-bearing: in the proof of Theorem 1, the divisor Q obtained from Eq. (3)-(4) has degree only ≥⌈n/2⌉ and can be special; Lemma 1(⇐) is then applied to Q. Since Lemma 6 and hence Theorem 2 invoke Theorem 1, the current proof depends on an unverified strengthening. I note that in Lemma 6 the subdivisor Q_i has degree ≥2g−1 and is therefore non-special, so that application could be repaired by using Orevkov's original non-special converse; but Theorem 1 as stated and Corollary 1 still require the stronger form. Please supply a proof of the strengthened converse, or modify the statement/proofs to avoid it.","section":"§2, Lemma 1 (footnote 1) and Theorem 1, Eq. (4)"}],"minor_comments":[{"comment":"The second binomial term counts tuples with sum 2g−2 and all entries even. Writing d_i=2e_i with e_i≥1 and ∑e_i=g−1 gives C(g−2,r−1), not C(g−1,r−1). For example, g=3,r=2 gives only (2,2), yet the displayed binomial equals 2. The bound is still finite, so this does not affect the main theorem, but the formula is incorrect.","section":"§3.1, Lemma 4, inequality (⋄)"},{"comment":"The assertion that d=d(1)+(0,k) cannot be written as a sum of elements of Sep(X) involving Sep_n(X) is not evident and as written appears false, since d∈Sep_n(X) by Lemma 2 (d(1) is the degree partition of a non-special separating divisor). The ratio argument that follows only excludes representations using elements of Sep_s(X). Please repair the argument or clarify the intended notion of generation.","section":"§3.2, Corollary 2 proof"},{"comment":"The phrase 'not all of the same sign' is ambiguous when some coefficients are zero. Please state explicitly that one chooses a relation with at least one positive and at least one negative coefficient (this can be achieved by subtracting a suitable multiple of the all-ones relation), so that β_s>0 and the set Γ in (3) is proper.","section":"§2, Theorem 1 proof, Eqs. (1)-(3)"},{"comment":"The sentence 'it suffices to obtain a uniform bound on the degrees of the minimal non-special divisors' should also recall that all elements of Sep_s(X) have |d|≤2g−2 by the proof of Lemma 4, so that for fixed g,r the set Sep_s(X) ranges over only finitely many subsets of N^r. This is implicit but should be stated.","section":"§4, Theorem 2 proof"},{"comment":"In 'We always equip the sets N r a N r 0' the symbol 'a' appears to be a typo for 'and'.","section":"§1, Notation"}],"recommendation":"major_revision","confidential_remarks":"The main theorem is likely correct and the proof strategy is sound, but the unproved strengthening of Orevkov's lemma must be addressed. The combinatorial miscount and the issue with Corollary 2 are secondary but should be fixed. I recommend major revision rather than rejection because the gap is local and the central finiteness argument can be repaired."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is the first finiteness theorem for separating semigroups of real curves of fixed genus, and the proof strategy — deletion lemma plus decomposition into a finite special part and finitely many cones — is fresh and mostly clean. I would send it to a referee, but I would not certify the main theorem as it stands, because the one genuinely load-bearing step is an unproved footnote assertion.\n\nWhat's new: Theorem 2 is not in the existing literature. Kummer–Shaw and Orevkov give explicit descriptions for M-curves, low genus, hyperelliptic curves, and plane quintics, but no finiteness for all genus g. Theorem 1 (the deletion lemma) is a nice linear-algebra argument using Abel–Jacobi: from a separating divisor of degree n, you get a proper separating subdivisor of degree at least n/2. Lemma 4's decomposition of Sep(X) into a finite set plus finitely many translated orthants is a good structural way to think about the semigroup, and Corollary 2's infinite-generation argument is elegant. The degree bound deg P_i ≤ 4g−3 is then straightforward.\n\nThe soft spot is exactly where the stress-test note points: Lemma 1 is stated as an iff, with the (⇐) direction attributed to Orevkov's Lemma 3.2. The footnote admits Orevkov's version assumed non-specialness and says the assumption 'can be removed,' but gives no proof. Theorem 1 applies that converse to the subdivisor Q, which need not be non-special. If the strengthening is false, Theorem 1, Corollary 1, and Lemma 6 all collapse. This is not a cosmetic gap; it is the load-bearing point. It may well be true — the Abel–Jacobi picture suggests a perturbative proof — but the paper does not supply one. A referee should ask for it before the main theorem is accepted.\n\nMinor things: the count for canonical degree partitions in Lemma 4 looks off by one — positive even r-tuples summing to 2g−2 should be binomial(g−2, r−1), not binomial(g−1, r−1) — but it does not affect finiteness. Also, the ⊔ notation in Lemma 4's proof is too strong if a single degree partition can be realized by both special and non-special divisors, which the paper itself raises in Question 1.\n\nFor whom: anyone working on real algebraic curves, separating morphisms, or the Kummer–Shaw invariant. The paper is short, clearly written, and the main idea is worth knowing. I'd send it to a serious referee with the explicit request to prove or source the strengthened converse of Orevkov's lemma. If the author fills that gap, the paper is solid.","headline":"First finiteness theorem for separating semigroups of real curves, with a genuinely nice deletion lemma — but it leans on an unproved strengthening of Orevkov's lemma that the author flags only in a footnote.","tokens_in":8451,"tokens_out":6012,"would_cite":false,"duration_ms":60880,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14H55","14P25","14H05"],"pacs":[],"model":"deepseek-v4-flash","headline":"For each genus, only finitely many separating semigroups exist","keywords":["separating semigroup","real algebraic curve","separating morphism","real curve divisors","Abel-Jacobi map","special divisor","finiteness theorem","gonality"],"falsifier":"Take a special separating divisor P of degree at least g+2 on a real curve of genus g and check whether P admits positive tangent vectors v_i at its points with ∑ω_k(v_i)=0 for all holomorphic one-forms ω_k. If some such divisor fails this condition, the paper's strengthened tangent-vector criterion is false, and the deletion theorem and the bound deg P_i ≤ 4g−3 no longer follow. Alternatively, exhibit a separating semigroup whose minimal non-special separating divisor has degree exceeding 4g−3.","tokens_in":7651,"feed_emoji":"🧮","tokens_out":5974,"duration_ms":50629,"temperature":0.7,"pith_summary":"The paper proves a finiteness theorem for separating semigroups of real algebraic curves: for a fixed genus g, there are only finitely many distinct separating semigroups among all real curves of that genus. This settles a structural question about how the degrees of separating morphisms to the projective line can vary. The proof shows that every separating semigroup is determined by two finite pieces of data—a finite set of special divisors and a finite set of minimal non-special separating divisors—and bounds the degrees of the latter by 4g−3. The key deletion theorem says that from any sufficiently large separating divisor one can remove at least half its points while preserving the separating property.","feed_headline":"For each genus, only finitely many separating semigroups exist","feed_subtitle":"Real curves of fixed genus organize their separating maps into a short, finite list.","key_machinery":"The load-bearing tool is the infinitesimal Abel-Jacobi criterion for separating divisors: a totally real divisor P = p_1 + ... + p_n is separating if and only if there are positive tangent vectors v_i at the points p_i (positive with respect to a complex orientation) such that ∑_i ω_k(v_i)=0 for every holomorphic one-form ω_k. This converts the separability condition into linear algebra over the reals. The other central mechanism is the deletion theorem: for n ≥ g+2, linear dependence among the vectors u_i=(ω_1(v_i),...,ω_g(v_i)) forces a proper non-empty sub-sum with positive coefficients involving at least half the points, so the corresponding subdivisor is again separating. This yields th","core_discovery":"The paper's main theorem, Theorem 2, states that for every non-negative integer g, the set of all separating semigroups of real curves of genus g is finite. To reach this, the paper proves a structural decomposition: the separating semigroup of any real curve is the union of a finite set of degree partitions coming from special separating divisors together with finitely many cofinal tails d(P_i) + N_0^r attached to minimal non-special separating divisors. The degrees of these minimal divisors are uniformly bounded by 4g−3. The bound follows from a deletion theorem: if a separating divisor has at least g+2 points, then some proper subdivisor containing at least half the points is still separa","pith_inferences":["If the strengthened tangent-vector criterion holds in full generality, the bound 4g−3 is probably far from optimal—the author notes known examples have much smaller degrees, so a sharper bound near 2g or even g+1 may hold.","The finiteness theorem suggests that separating semigroups are stable invariants that could, in principle, be enumerated by scanning finitely many divisor classes; a computer search for small g would test how close the 4g−3 bound is to the true maximum.","The dependence on the strengthened converse of the tangent-vector criterion is the fragile point: if that converse fails for special divisors, the deletion theorem may fail exactly at the degrees needed and the main theorem would not follow from the given arguments.","A similar finiteness statement may hold for separating semigroups of real curves with marked points or for separating maps to higher-genus targets, since the underlying tangent-vector linear algebra is the same."],"forward_implications":["Every separating curve of genus g admits a separating morphism of degree at most g+1, recovering the classical sharp bound on separating gonality.","For real curves with at least two real components, the separating semigroup is not finitely generated; only the finite special-degree part and finitely many additive tails exist.","The uniform degree bound 4g−3 places all minimal non-special separating divisors inside a finite box, so the full separating semigroup can in principle be found by a finite search.","For fixed genus g and number of real components r, the number of possible separating semigroups is bounded by an explicit expression in g and r, namely (4g−3)^r times a combinatorial factor for special partitions.","Known classifications of separating semigroups for small genera (up to 4) are consistent with the new uniform finiteness statement."],"fun_headline_variants":["For each genus, finitely many separating semigroups","Fixed genus: finite separating semigroups","Each genus yields finite separating semigroups","Separating semigroups: finite list per genus","For fixed genus, separating semigroups are finite"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The argument rests on the assertion that the tangent-vector criterion is an if-and-only-if even for special divisors—the paper cites this strengthened converse as removable but gives no proof; if it is false, the deletion theorem's output subdivisor may not be separating and the degree bound 4g−3 collapses.","fun_headline_variants_meta":{"raw":{"variants":["For each genus, finitely many separating semigroups","Fixed genus: finite separating semigroups","Each genus yields finite separating semigroups","Separating semigroups: finite list per genus","For fixed genus, separating semigroups are finite"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001252,"raw_usage":{"total_tokens":4955,"prompt_tokens":718,"completion_tokens":4237,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":462,"completion_tokens_details":{"reasoning_tokens":4164}},"tokens_in":462,"tokens_out":4237,"duration_ms":26793,"temperature":1.0,"reasoning_tokens":4164,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T20:42:14.306950+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a special separating divisor P of degree at least g+2 on a real curve of genus g and check whether P admits positive tangent vectors v_i at its points with ∑ω_k(v_i)=0 for all holomorphic one-forms ω_k. If some such divisor fails this condition, the paper's strengthened tangent-vector criterion is false, and the deletion theorem and the bound deg P_i ≤ 4g−3 no longer follow. Alternatively, exhibit a separating semigroup whose minimal non-special separating divisor has degree exceeding 4g−3.","supporting_citations":[],"review_version":1}