{"id":"3a163123-a79b-42c0-9825-3d3a505ff5f8","arxiv_id":"2608.07212","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Every reduced real plane curve singularity is shown to admit a real morsification, via a new trace map built from nodal smoothings of normalization disks.","lead":"Every reduced real plane curve singularity admits a real morsification, a real deformation whose nearby fibers have only simple real crossings. This settles conjectures of Leviant and Shustin and of Fomin, Pylyavskyy, Shustin, and Thurston, using a new trace map construction for pairs of complex conjugate branches.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified; the flagged separation assertion follows from the definition of last common real infinitely near point.","rationale":"The central claim is Theorem A, asserting existence of a real morsification for every reduced real plane curve germ. The proof is long but structured: the trace map handles conjugate pairs with distinct nonreal tangents; packets at last common real infinitely near points are shown to satisfy this condition; translations and contractions assemble the divide; and an Euler characteristic comparison gives the matching upper bound. I examined the reader's flagged assumption in detail. The assertion that at a last common real infinitely near point the two tangent lines of a conjugate pair are distinct and nonreal is in fact a direct consequence of the meaning of 'last common real infinitely near point': if the tangents were equal, the tangent line would be invariant under the real structure, hence real, and the strict transforms would share the corresponding real point on the exceptional divisor after the next blowup, contradicting minimality. Thus the trace-map construction never needs to handle the self-tangency case of Example 5.27. The other potential soft spots—the finite base-change/normalization argument and the Euler characteristic computation—are sketched but standard: the Euler characteristic of the normalization (rebr) minus twice the number of nodes equals the Milnor fiber Euler characteristic, giving n = delta - imbr. No concrete error or missing essential proof was identified, so the verdict should remain unchanged.","tokens_in":32074,"tokens_out":36443,"duration_ms":363077,"concrete_test":"Verify the separation assertion by fully resolving a conjugate pair with a common real tangent, e.g. the pair of Example 5.27, until the strict transforms separate, and check at the final common real infinitely near point that the two tangent lines form a distinct nonreal conjugate pair and that the strict transforms meet the exceptional divisor at two distinct nonreal points. If any resolution produced a common real point after the claimed last one, the descent step would fail.","verdict_should_be":"UNCHANGED","load_bearing_attack":"No significant objection identified. The reader's main concern—that Theorem 6.8, step 3, needs the two tangent lines of each conjugate pair at a last common real infinitely near point to be distinct and nonreal—is not load-bearing, because the assertion is forced by the resolution setup. If such a pair had equal tangent lines, the line would be real, and after blowing up the point the strict transforms would meet at the real point of the exceptional divisor corresponding to that common real tangent, contradicting the choice of c as the last common real infinitely near point. Example 5.27 only demonstrates that the trace-map circle construction fails in exactly this excluded case. The remaining summarized steps (finite base changes and the Euler characteristic comparison) are standard and reproduce n = delta(C,0) - imbr(C,0). I found no internal inconsistency or unjustified assumption that would change the verdict.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves Theorem A: every reduced real plane curve germ (C,0) admits a real morsification, i.e., a real nodal deformation whose nearby fibers have exactly delta(C,0) - imbr(C,0) real hyperbolic nodes and no other singularities. The proof introduces a trace map on a smoothing of two normalization disks. For a conjugate pair of nonreal branches with distinct tangents, the trace map gives an explicit predivide whose double-point count is computed from the Puiseux data; for a real branch it recovers Gusein-Zade's Chebyshev construction. The global construction assembles these local predivides along a real embedded resolution using translations and contractions, and the final node count is verified by an Euler characteristic comparison with the Milnor fiber.","tokens_in":32232,"tokens_out":27196,"duration_ms":227395,"significance":"If correct, Theorem A settles a question of A'Campo and Gusein-Zade and the later conjectures of Leviant--Shustin and of Fomin--Pylyavskyy--Shustin--Thurston. The trace map is a new and explicit mechanism for producing divides for conjugate pairs, and the paper gives a worked example of the previously open Leviant--Shustin case as well as a complete seven-branch example. The main analytic estimates are presented in detail and the final count is checked independently by the Euler characteristic, so the result does not rest on fitted parameters or circular reasoning. The argument is long, but the reliance on standard tools is explicit.","major_comments":[],"minor_comments":[{"comment":"Several overlines appear to be missing from the typesetting. In Lemma 6.1 the sentence \"Since L̸= L, Q_i and Q_j have distinct tangent lines, as do Q_i and Q_j\" is inconsistent with the hypothesis that both Q_i and Q_j have tangent L; the intended statement is that Q_i and \\bar Q_j have distinct tangents, and likewise \\bar Q_i and Q_j. Similarly, (5.2) should display the real structure as (p,q) mapping to (\\bar q, \\bar p), with the real plane given by q=\\bar p. These are notation issues, but they should be corrected to avoid confusion.","section":"Lemma 6.1 and Section 5"},{"comment":"The assertion that at a last common real infinitely near point c the two tangent lines of each conjugate pair are distinct and nonreal is stated in one sentence. I verified the argument: if the tangents coincided, the common tangent would be real and, after blowing up c, the strict transforms would meet at the real point of the exceptional divisor corresponding to that tangent, contradicting the choice of c. I recommend adding this one-line explanation explicitly, since Lemma 6.7 depends on it.","section":"Theorem 6.8, descent step"},{"comment":"The computation that yields \\sum_{q \\in \\operatorname{Sing}(Y_\\lambda)} \\delta(Y_\\lambda,q) = \\delta(C,0)-\\operatorname{imbr}(C,0) is summarized in a single sentence. A short derivation, using \\chi(Y_\\lambda)=\\chi(S_\\lambda)-n for the n nodes and \\chi(\\operatorname{Milnor fiber})=\\chi(S_\\lambda)-2n, would help the reader check the factor of one half, which is load-bearing for the final count.","section":"Theorem 6.8, Euler characteristic comparison"},{"comment":"There are several typographical errors that should be fixed: \"Aknowledgements\", \"contracions\", \"with with\" in Section 6, \"all the and real blow ups\" in Section 7, and the reference to \"step 6\" for the final perturbation in the proof of Theorem 6.8, which appears to mean step 4. In Example 5.29 the displayed formula for f_r(\\theta) omits the overline on the second term; it should be w_\\theta^2 + \\overline{w_\\theta}^3.","section":"Throughout"}],"recommendation":"minor_revision","confidential_remarks":"This is a strong paper with a significant main theorem. I found no load-bearing error; the one-sentence separation assertion in Theorem 6.8 is correct as written, and the Euler characteristic check is sound. The minor issues are presentational. The editor may wish to confirm that the cited sources [Cas15] and [APC25] are accessible, since the parametrized translation step is cited to them."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is the first proof that every reduced real plane curve singularity admits a real morsification, and the main construction is genuinely new. The trace map turns a smoothing of two normalization disks into a real predivide, and for a conjugate pair with distinct nonreal tangents it gives an explicit divide formula in Puiseux data. That covers exactly the cases Leviant–Shustin left open, including their simplest example. I checked the node counts in Proposition 5.19 and Lemmas 6.1 and 6.7 with some care; they are internally consistent, and the final count matches the independent Euler characteristic bound. The paper is also honest about its limits: Example 5.27 shows the trace-map circle is not a predivide when a conjugate pair shares a real tangent, and Example 5.29 shows the count depends on choosing a Newton–Puiseux parametrization. Neither undermines Theorem A because the global proof handles shared tangents through the resolution construction.\n\nThe soft spots are real but minor. The separation assertion in Theorem 6.8, step 3 is stated in one sentence (“otherwise a common tangent would be real…”). The stress-test note is right: if two conjugate branches had the same tangent at their last common real infinitely near point, the strict transforms would still meet in the exceptional divisor, so that point would not be last. Still, the paper would be easier to referee if this were spelled out. The finite base-change step and the Euler characteristic comparison are also summarized, with references to Wall and to Castellini/Leviant–Shustin; they are standard in this context, and the final explicit seven-branch example in Section 7 gives me additional confidence that the assembly works.\n\nCitation practice looks fine. The paper cites Leviant–Shustin, Fomin–Pylyavskyy–Shustin–Thurston, A’Campo, Gusein-Zade, and Wall throughout; the self-citations are expository. No fitted constants and no circularity: the double-point count comes from the Puiseux series and Wall’s semigroup formula, and the upper bound is an independent Euler characteristic computation.\n\nWho this is for: singularity theorists working on real morsifications, divides, and quiver mutation. It resolves the existence question and makes the FPST mutation conjecture meaningful in the conjugate-pair case. This deserves a serious referee. I would send it to review, expecting minor revision to expand the separation argument and clarify the base changes. I would cite it if I continued working on divides.","headline":"Proves the long-open real morsification conjecture with a genuinely new trace-map construction, and the core counts hold up under scrutiny.","tokens_in":32742,"tokens_out":2270,"would_cite":true,"duration_ms":37950,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14B05","14H20","14P05","32S25"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that every reduced real plane curve germ, in any real form, admits a real nodal deformation with exactly $\\delta(C,0)-\\operatorname{imbr}(C,0)$ real hyperbolic nodes and no other singularities.","keywords":["real morsification","plane curve singularity","divides","trace map","conjugate branches","Newton–Puiseux parametrization","nodal deformation","delta invariant"],"falsifier":"Run the full resolution descent on the seven-branch example of Section 7 and check that the final polynomial has exactly 42 real hyperbolic nodes for all sufficiently small positive parameters; any other count would show the $\\delta-\\operatorname{imbr}$ bound is not attained. A complementary check is to search for a reduced real germ whose resolution has a last common real infinitely near point where a conjugate pair shares a real tangent: there Lemma 6.7 would produce a self-tangency, as in Example 5.27, and the descent count would fail.","tokens_in":31891,"feed_emoji":"🪢","tokens_out":13034,"duration_ms":113282,"temperature":0.7,"pith_summary":"The paper proves that every reduced real plane curve germ, in any prescribed real form, admits a real morsification: a real deformation whose nearby fibers have exactly $\\delta(C,0)-\\operatorname{imbr}(C,0)$ real hyperbolic nodes and no other singularities. This was the missing piece left by earlier constructions, which worked only when all branches were real. The new engine is the trace map, which produces an explicit real circle from the nodal smoothing of two normalization disks; for a conjugate pair of nonreal branches, the self-intersections of this circle are computed directly from Puiseux data and realize the required number of double points. The proof assembles these circles with real arcs and translation steps through a resolution descent, and an Euler characteristic comparison with the Milnor fiber supplies the matching upper bound. As a result, every real form has a divide, and the earlier conjectural statements about real morsifications become theorems.","feed_headline":"Every real plane curve singularity admits a real morsification","feed_subtitle":"A trace map turns nodal smoothings of normalization disks into real circles, producing divides for every real form.","key_machinery":"The trace map is the map $\\Phi_\\gamma(u,v,s)=\\gamma_Q(u)+\\gamma_{\\bar Q}(v)$ on the smoothing $A=\\{uv=s\\}$ of two normalization disks. Its key property is equivariance with respect to the real structure $(u,v,s)\\mapsto(v,u,s)$; the fixed circle $(u,v)=(re^{i\\theta},re^{-i\\theta})$ maps into the real plane and, for small $r$, is a predivide. The counting engine is the Puiseux-data calculation: with $\\gamma_Q(u)=(u^m,\\varphi(u))$ and $\\kappa(\\zeta)=\\operatorname{ord}_u(\\varphi(u)-\\varphi(\\zeta u))$ for $\\zeta^m=1$, $\\zeta\\neq1$, the number of double points of the circle is $\\sum_{\\zeta\\neq1}(m+\\kappa(\\zeta))=2\\delta(Q,0)+m^2-1$, so the pair $Q\\cup\\bar Q$ contributes $\\delta(Q\\cup\\bar Q,0)-1$. Lemma 6.7 extends this to packets of several conjugate pairs at a real infinitely near point and to their intersections with real arcs, which is what makes the global resolution descent work.","core_discovery":"The central claim, Theorem A, is that every reduced real plane curve germ $(C,0)$ admits a real morsification: a real nodal deformation such that, for every sufficiently small positive parameter, the fiber has exactly $\\delta(C,0)-\\operatorname{imbr}(C,0)$ real hyperbolic nodes and no other singularities. The obstruction to previous approaches was a conjugate pair $Q,\\bar Q$ of nonreal branches. The paper removes it with the trace map on the smoothing $A=\\{uv=s\\}$ of two normalization disks, $\\Phi(u,v,s)=\\gamma_Q(u)+\\gamma_{\\bar Q}(v)$. The real structure sends $(u,v,s)$ to $(v,u,s)$, so the fixed locus of the $s=r^2$ fiber is the circle $(re^{i\\theta},re^{-i\\theta})$, and its image is a real predivide. For distinct tangents, the double points are counted exactly and equal $\\delta(Q\\cup\\bar Q,0)-1$, organized in root-of-unity sectors determined by the Puiseux exponents; for real branches the same construction, after a quotient, recovers the classical Chebyshev parametrizations. Section 6 combines these trace-map circles with real arcs, translations, and contractions along the resolution, and the Euler characteristic of the Milnor fiber forces the node count up to the upper bound $\\delta-\\operatorname{imbr}$.","pith_inferences":["The explicit Puiseux formula for the trace-map circle suggests a direct algorithm for drawing divides from the characteristic exponents alone, bypassing the full resolution; the paper does not draw this algorithmic conclusion.","When a conjugate pair shares a real tangent, the circle map develops a self-tangency rather than a predivide; a limiting or perturbed version of the trace map might still recover the correct count, but the paper leaves that case to the resolution descent.","Because a divide encodes the singularity link and monodromy, the new divides for non-totally-real forms may give explicit vanishing-cycle data for those real forms; the paper does not compute such data.","The seven-branch example's explicit polynomial is a concrete test case for numerical singularity software: the predicted 42 real hyperbolic nodes can be checked for small positive parameter values."],"forward_implications":["Every prescribed real form of a complex plane curve singularity now has a real morsification, so the real-form restriction is no longer an obstacle.","The divide of the real locus is explicitly assembled from trace-map circles and translated arcs; for conjugate pairs with distinct tangents it is computable from Puiseux data.","The maximum number $\\delta(C,0)-\\operatorname{imbr}(C,0)$ of real hyperbolic nodes allowed by the earlier bound is attained for every reduced real germ.","For any two real forms of the same complex singularity, both now admit real morsifications, so the statement that their associated quivers are mutation equivalent can be tested across all real forms.","The construction can be made explicit enough to produce a polynomial defining the morsification, as in the worked seven-branch example."],"supporting_citations":[{"why":"Supplies the upper bound $\\delta-\\operatorname{imbr}$ and the criterion that attaining it with no other singularities characterizes real morsifications; also gives the example left open and the descent pattern generalized here.","marker":"[LS18]"},{"why":"Provides the translation-and-contraction method and the divide notion that Section 6 extends to conjugate pairs.","marker":"[A'C75]"},{"why":"Gives the Chebyshev-polynomial morsification of real branches that the trace map recovers in the real-branch case.","marker":"[GZ74a]"},{"why":"Supplies the standard formulas for $\\delta$, intersection multiplicities, and blowup behavior used to compute the node count and the Euler characteristic bound.","marker":"[Wal04]"},{"why":"The proper mapping theorem turns the parametrized descent construction into an actual holomorphic equation, i.e. a real deformation.","marker":"[Rem56]"},{"why":"Provides the parametrized form of the translations used before each contraction in the descent.","marker":"[Cas15]"},{"why":"Formulates the conjecture whose standing hypothesis, existence of real morsifications for all real forms, Theorem A supplies.","marker":"[FPST22]"}],"fun_headline_variants":["Trace map settles real morsification conjecture","All real plane curves admit real morsifications","Trace map overcomes last hurdle: real morsifications","Real morsification for every curve via trace map"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The node count rests on the assertion that at every last common real infinitely near point of a conjugate pair, the two tangent lines are distinct and nonreal; the proof states this follows because a common real tangent would create another common real infinitely near point, but the separation step is not proved in detail.","fun_headline_variants_meta":{"raw":{"variants":["Trace map settles real morsification conjecture","All real plane curves admit real morsifications","Trace map overcomes last hurdle: real morsifications","Real morsification for every curve via trace map"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000404,"raw_usage":{"total_tokens":2126,"prompt_tokens":989,"completion_tokens":1137,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":605,"completion_tokens_details":{"reasoning_tokens":1076}},"tokens_in":605,"tokens_out":1137,"duration_ms":11773,"temperature":1.0,"reasoning_tokens":1076,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T12:39:17.210995+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the full resolution descent on the seven-branch example of Section 7 and check that the final polynomial has exactly 42 real hyperbolic nodes for all sufficiently small positive parameters; any other count would show the $\\delta-\\operatorname{imbr}$ bound is not attained. A complementary check is to search for a reduced real germ whose resolution has a last common real infinitely near point where a conjugate pair shares a real tangent: there Lemma 6.7 would produce a self-tangency, as in Example 5.27, and the descent count would fail.","supporting_citations":[],"review_version":1}