{"id":"203c67c7-8987-4bff-bc24-d8bbd695aa35","arxiv_id":"2508.18543","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":1,"one_line_summary":"Two errors in the proof of the main results of a paper by Jang, So, and Marotta are identified and corrected while keeping the main results true.","lead":"A short math note finds two errors in a 2017 proof about baby Mandelbrot sets in a family of rational maps, and proposes corrected constructions to rescue the original conclusions. It is a correction note for specialists in complex dynamics, not a new discovery.","discovery_kind":"replication","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The correction hinges on the unproved inclusion bUλ⊂Vλ; without it, bU'λ may not lie in V'λ and polynomial-likeness is not established.","rationale":"The reader's weakest assumption is precisely the unproved inclusion bUλ⊂Vλ. I agree. The stress-test pass did not find a separate weakness; this unproved inclusion is the load-bearing point. The paper's own text flags it implicitly by asserting it without demonstration. The recommended verdict remains CONDITIONAL: acceptance should require a proof (or computational verification with error bounds) of bUλ⊂Vλ. I did not find grounds to reject outright, because the inclusion may be true; it is simply unsupported. Thus the reader's verdict is unchanged.","tokens_in":5772,"tokens_out":18102,"duration_ms":181068,"concrete_test":"For a representative parameter with n̸=d (e.g., n=3, d=2, λ on ∂W with ψ=0), compute Fλ(V'λ) by image-sampling a fine grid in V'λ (or by tracing the boundary curves bγn, bγd and the ray segments), and test whether the open sector bUλ={|z|<2+ε, θ2<Arg z<θ1} is contained in the computed image. If any point of bUλ—especially near the origin and near the inner boundary—has no preimage in V'λ, the asserted inclusion is false and the correction does not establish polynomial-likeness. Repeat for ψ=±π/(n−1) to cover ∂W.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 4 defines V'λ and Vλ=Fλ(V'λ), then defines bUλ as the sector {|z|<2+ε, θ2<Arg z<θ1} and asserts 'Since bUλ⊂Vλ, we have by choice of component that bU'λ⊂V'λ.' This inclusion is never proved, and it is the step that makes the correction work. Without it, the chosen preimage component bU'λ could cross the boundary of V'λ, so the restriction Fλ:bU'λ→bUλ is not known to be proper of degree two; Proposition 2 fails. Note that Proposition 1 proves the opposite-direction inclusion V'λ⊂bUλ (domain inside target), which does not imply bUλ⊂Vλ. The containment depends on whether the arcs Fλ(bγn) and Fλ(bγd) of bΓ together cover the entire bΓ-arc of bUλ and whether the images of the two ray-segments of ∂V'λ lie exactly on the boundary rays of bUλ. None of this is checked in the paper. Since the main claim that the original results remain true relies on Proposition 2 and Proposition 3, the correction is conditionally supported at best.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims to identify two errors in [5]'s proof that the maps Fλ(z)=z^n+λ/z^d form a polynomial-like family of degree two and satisfy the Douady-Hubbard winding condition. The first alleged error is that the boundary ray segments of U'λ cannot map 2-to-1 onto the piecewise linear boundary of Uλ because Fλ is conformal away from critical points and the relevant rays meet at the origin at an angle that is not π when n≠d. The second alleged error is that, for λ on the outer circular boundary of W, the critical value vλ lies on Γ, which is part of the boundary of the open set Uλ, violating the hypothesis that vλ lie in Uλ\\U'λ. The paper then proposes new regions bUλ (a Pac-Man sector with outer radius 2+ε and restricted angle bounds) and bU'λ (a chosen preimage component), and asserts that these repair both defects. It concludes that Proposition 5.1 and Lemma 5.2 of [5], and hence the existence of n−1 baby Mandelbrot sets, remain true.","tokens_in":6057,"tokens_out":13725,"duration_ms":142445,"significance":"The diagnostic part of the paper is valuable and largely convincing: the conformality argument showing that the original 2-to-1 boundary mapping is impossible is sound, and the observation that the original winding argument places vλ on the topological boundary of an open domain is a genuine gap. If the proposed fix were fully proved, the note would be a useful correction to the literature. However, the paper's central claim—that the new regions make Fλ polynomial-like of degree two and satisfy the winding condition—is not established in the manuscript. The proof relies on asserted containments and on the very degree-two property to be proved, and Proposition 3 imports the path of vλ from the very proof being corrected. The significance is therefore conditional on substantial additional argument.","major_comments":[{"comment":"The inclusion bUλ⊂Vλ is asserted without proof, and this inclusion is load-bearing for the whole construction. From it the paper infers that bU'λ⊂V'λ and then, using Proposition 1, that bU'λ⊂bUλ. Proposition 1 proves only the opposite inclusion V'λ⊂bUλ, so it cannot justify the assertion. One needs to prove that the image of ∂V'λ surrounds bUλ: namely, that the arcs Fλ(bγn) and Fλ(bγd) together cover the bΓ-arc between the two boundary rays of bUλ, and that the images of the two ray segments of ∂V'λ lie exactly on the boundary rays of bUλ. Without this, the chosen preimage component bU'λ could cross ∂V'λ, and the restricted map Fλ:bU'λ→bUλ is not known to be proper.","section":"Section 4, paragraph after Proposition 1"},{"comment":"The sentence 'since bUλ is the subset of Vλ onto which Fλ maps V'λ in a 2-to-1 fashion' is circular: this is precisely the polynomial-like degree-two statement that Proposition 2 is supposed to prove. The proof must instead derive the 2-to-1 property from a verified boundary covering or a degree computation, e.g., via the argument principle, and must show that bU'λ is a bounded simply connected domain and that Fλ:bU'λ→bUλ is proper. The final sentence of the proof also uses the unhatted symbols U'λ and Uλ, while the proposition concerns bU'λ and bUλ.","section":"Section 4, Proposition 2 proof"},{"comment":"The angle inequalities in Proposition 1 are not true for the stated range |ψ|≤π/(n−1). For example, take n=3, d=2, so m=5, and choose ψ=−π/2, which satisfies |ψ|≤π/(n−1)=π/2. In the case n>d the proof claims ψ+π/m < nψ+dπ/m, but here ψ+π/m=−3π/10 and nψ+dπ/m=−11π/10, so the claimed inequality fails. Consequently V'λ⊄bUλ for this point, and Proposition 1 as stated is false. Since Proposition 1 is used to conclude bU'λ⊂bUλ and is also relied on in the angle estimate in Proposition 3, the admissible range of ψ for which the construction works must be redetermined and stated correctly.","section":"Section 4, Proposition 1"},{"comment":"The proof of the winding condition is incomplete. The assertion that 'the path of the critical value vλ remains the same as in the proof of Lemma 5.2' imports the very path from the proof that the paper is correcting, so it cannot be taken as an established fact. More importantly, on the arcs where |vλ|=2 the claim that vλ∈bUλ∖bU'λ is not proved: since Γ lies strictly inside bUλ after the ε enlargement, membership in bUλ is automatic, but vλ∉bU'λ requires a separate argument. On the ray segments, the displayed inequalities compare only arguments and do not compare moduli, nor do they establish strict exclusion or handle boundary equality; and showing vλ∈bUλ∖bU'λ pointwise on ∂W does not by itself show that the loop winds once around the exterior of bU'λ. The winding number around bU'λ must be computed explicitly.","section":"Section 4, Proposition 3"}],"minor_comments":[{"comment":"There are typographical errors: 'Propostition' in Section 4 and 'symetrically' in Section 2 should be corrected.","section":"Throughout"},{"comment":"The quotation of the Douady-Hubbard definition appears incomplete as typeset: after the first bullet the second defining condition (properness, or relative compactness of U′ in U) is missing, which makes the statement of the theorem being used unclear.","section":"Section 2, Definition 1"},{"comment":"The notation bUλ−bU'λ and bUλ−bU'λ should be set difference, e.g., bUλ∖bU'λ, to avoid confusion with pointwise subtraction.","section":"Section 4, Proposition 2 proof"},{"comment":"The proof of Proposition 3 says that if Argλ=π/(n−1), then for any z∈bU'λ one has Argz≤ψ+π/m. This estimate depends on the unproved inclusion bU'λ⊂V'λ and on the angle bounds of V'λ, so the assertion should be restated after those facts are established.","section":"Section 4, Proposition 3"}],"recommendation":"major_revision","confidential_remarks":"The paper is a corrigendum-style note whose diagnostic discussion of the two errors in [5] appears sound and worth publishing if the fix can be completed. As it stands, the proof of the corrected construction has load-bearing gaps: the inclusion bUλ⊂Vλ, the 2-to-1 property, and the winding argument are all asserted rather than derived, and Proposition 1 appears to be false on the stated parameter range. I would ask for a revised version with complete proofs of those points before recommending publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague — short version: this note convincingly identifies two genuine gaps in [5], but the repair is incomplete. The geometric objection to Proposition 5.1 is sound: a conformal map cannot send a smooth ray segment 2-to-1 onto a corner at the origin unless the angle matches, and for n≠d it does not. That part is precise and worth knowing. The observation that Uλ must be open, so the critical value lying on Γ violates the winding condition, is also correct.\n\nWhat is new: the two error identifications are specific and previously unnoticed, and the proposed modified regions (enlarged outer circle, narrowed angle) are a reasonable direction. The note is honest about the n=d special case and correctly distinguishes [5] from the Boyd–Mitchell and Boyd–Hoeppner work where n=d.\n\nThe soft spots are where the reader's report puts them, and they are load-bearing. In Section 4, after defining V′λ and Vλ, the paper states 'Since bUλ ⊂ Vλ' — this is never proved. Proposition 1 proves V′λ ⊂ bUλ, which is the opposite direction and does not imply bUλ ⊂ Vλ. Without that inclusion, the choice of component bU′λ ⊂ V′λ is unjustified, and Fλ: bU′λ → bUλ is not shown to be proper of degree two. Moreover, the sentence 'since bUλ is the subset of Vλ onto which Fλ maps V′λ in a 2-to-1 fashion' is exactly the polynomial-likeness statement under proof; that is circular. Proposition 3 also inherits the old winding path from [5] without rechecking it against the new bUλ, so the second correction is not self-contained.\n\nFor the intended audience — complex dynamicists working on baby Mandelbrot sets and polynomial-like maps — the error report is useful, but the fix is not yet rigorous. I would send it to a serious referee, asking for a proof of bUλ ⊂ Vλ and a non-circular degree-2 argument. If those can be supplied, the note becomes a valuable correction to the literature.","headline":"Real errors found in Jang–So–Marotta, but the proposed fix is not yet proved: the key inclusion bUλ⊂Vλ is asserted, and Proposition 2 is circular.","tokens_in":6538,"tokens_out":3203,"would_cite":false,"duration_ms":31109,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["37F10","37F45"],"pacs":[],"model":"deepseek-v4-flash","headline":"Two proof errors fixed, baby Mandelbrot sets survive","keywords":["polynomial-like maps","baby Mandelbrot sets","rational maps","winding condition","connectedness locus","complex dynamics","prepoles","Mandelbrot set"],"falsifier":"Take a concrete pair such as $(n,d)=(3,2)$, choose $\\lambda$ with $\\operatorname{Arg}\\lambda=\\pi/(n-1)$ and $|\\lambda|=\\frac{n}{d}\\left(\\frac{2d}{m}\\right)^{m/n}$, and compute numerically whether every boundary point of $\\widehat U_\\lambda$ lies inside $V_\\lambda=F_\\lambda(V'_\\lambda)$. One boundary point of $\\widehat U_\\lambda$ escaping $V_\\lambda$ would falsify the unproved inclusion and therefore Proposition 2.","tokens_in":5588,"feed_emoji":"🌀","tokens_out":8330,"duration_ms":80123,"temperature":0.7,"pith_summary":"This note corrects two nontrivial errors in the proof that the rational family $F_\\lambda(z)=z^n+\\lambda/z^d$ contains $n-1$ baby Mandelbrot sets. The first error concerns the claim that the boundary of the preimage region maps 2-to-1 onto the boundary of the image region; the note shows this mapping is actually 1-to-1 and passes through the origin, so the original regions do not give a polynomial-like family. The second error is that, on part of the parameter boundary, the critical value lands on the boundary of the open region where it must lie, violating the winding condition. The note repairs both by redefining the image region with a slightly larger outer radius and narrower angle bounds, and proves that with these corrected regions the polynomial-like and winding conditions hold. A sympathetic reader should care because the existence of $n-1$ baby Mandelbrot sets is exactly what the original paper promised, and this correction shows that promise can be kept.","feed_headline":"Two proof errors fixed, baby Mandelbrot sets survive","feed_subtitle":"A repaired construction keeps the promised n−1 baby Mandelbrot sets for a family of rational maps","key_machinery":"The load-bearing object is the corrected Pac-Man region $\\widehat U_\\lambda$, defined by $|z|<2+\\varepsilon$ and angle bounds $\\theta_2<\\operatorname{Arg} z<\\theta_1$, with $\\theta_1,\\theta_2$ chosen from the arguments $n\\psi\\pm d\\pi/m$ of the image sector, together with its preimage component $\\widehat U'_\\lambda$ inside $V'_\\lambda$. The slightly enlarged outer boundary $\\widehat\\Gamma$ of radius $2+\\varepsilon$ ensures the critical value on $\\Gamma$ lies inside the open region rather than on its boundary, while the restricted angle bounds make $\\widehat U'_\\lambda\\subset\\widehat U_\\lambda$ available and restore the 2-to-1 covering needed for polynomial-like degree two.","core_discovery":"The central discovery is that the original regions $U'_\\lambda$ and $U_\\lambda$ are the wrong shapes for the conclusion, not the conclusion itself. With $n\\ne d$, the ray segments of $\\partial U'_\\lambda$ contain prepoles and therefore map to line segments through the origin; conformality forces these images to cross at the origin, so they cannot cover the two non-parallel boundary rays of $U_\\lambda$ in a 2-to-1 fashion. Replacing $U_\\lambda$ by a Pac-Man shaped region $\\widehat U_\\lambda$ whose outer radius is $2+\\varepsilon$ and whose angle bounds are taken from the image sector $V_\\lambda$, and taking $\\widehat U'_\\lambda$ as the corresponding preimage component, restores the 2-to-1 property. The note then proves that the critical value lies in $\\widehat U_\\lambda\\setminus\\widehat U'_\\lambda$ for every $\\lambda\\in\\partial W$ and winds once around $\\widehat U'_\\lambda$, so the polynomial-like straightening theorem applies and the original proposition and lemma are true.","pith_inferences":["A natural extension, not explored here, is that the same boundary-touching failure would appear in any polynomial-like construction whose critical value lands on the outer boundary by design; the epsilon-enlargement repair is generic but must be re-proved for each family.","One testable next step is to compute explicit bounds for $\\varepsilon$ in terms of $n$ and $d$; the proof only asserts existence of a sufficiently small $\\varepsilon$, so a quantitative version would make the correction easier to verify and reuse.","A reader might expect the corrected proof to transfer to nearby families with an added constant, provided the prepole-ray geometry and the non-parallel condition $n\\ne d$ are preserved; this note does not make that extension."],"forward_implications":["For every $\\lambda\\in W$, the corrected family $F_\\lambda\\colon \\widehat U'_\\lambda\\to \\widehat U_\\lambda$ is polynomial-like of degree two, so the straightening theorem applies to it.","For every $\\lambda\\in\\partial W$, the critical value stays in $\\widehat U_\\lambda\\setminus\\widehat U'_\\lambda$ and winds once around $\\widehat U'_\\lambda$, satisfying the winding condition.","The original claims of Proposition 5.1 and Lemma 5.2 in [5] are true as stated: the family is polynomial-like of degree two and there are $n-1$ baby Mandelbrot sets in the connectedness locus.","When $n=d$, no correction is needed; the earlier result already covers that case."],"supporting_citations":[{"why":"Supplies the original construction of $U_\\lambda$, $U'_\\lambda$, Proposition 5.1, and Lemma 5.2 that this note corrects.","marker":"[5]"},{"why":"Provides the polynomial-like map theorem and the winding condition that the corrected regions must satisfy.","marker":"[4]"},{"why":"Establishes the equal-exponent case $n=d$, which the note excludes and uses as a comparison boundary for its correction.","marker":"[3]"}],"fun_headline_variants":["Pac-Man fix restores 2-to-1 map, baby Mandelbrot sets safe","Proof error corrected, baby Mandelbrot sets survive","JSM proof patched, baby Mandelbrot sets remain true","Wrong regions fixed, baby Mandelbrot sets intact","2-to-1 mapping saved, baby Mandelbrot sets preserved"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire repair rests on the assertion, used without proof, that the new region $\\widehat U_\\lambda$ is contained in the image region $V_\\lambda$; if that inclusion fails, the chosen preimage component $\\widehat U'_\\lambda$ need not lie inside $V'_\\lambda$, and the 2-to-1 polynomial-like argument collapses.","fun_headline_variants_meta":{"raw":{"variants":["Pac-Man fix restores 2-to-1 map, baby Mandelbrot sets safe","Proof error corrected, baby Mandelbrot sets survive","JSM proof patched, baby Mandelbrot sets remain true","Wrong regions fixed, baby Mandelbrot sets intact","2-to-1 mapping saved, baby Mandelbrot sets preserved"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000211,"raw_usage":{"total_tokens":1349,"prompt_tokens":817,"completion_tokens":532,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":433,"completion_tokens_details":{"reasoning_tokens":436}},"tokens_in":433,"tokens_out":532,"duration_ms":5348,"temperature":1.0,"reasoning_tokens":436,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T16:58:25.086723+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a concrete pair such as $(n,d)=(3,2)$, choose $\\lambda$ with $\\operatorname{Arg}\\lambda=\\pi/(n-1)$ and $|\\lambda|=\\frac{n}{d}\\left(\\frac{2d}{m}\\right)^{m/n}$, and compute numerically whether every boundary point of $\\widehat U_\\lambda$ lies inside $V_\\lambda=F_\\lambda(V'_\\lambda)$. One boundary point of $\\widehat U_\\lambda$ escaping $V_\\lambda$ would falsify the unproved inclusion and therefore Proposition 2.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the original construction of $U_\\lambda$, $U'_\\lambda$, Proposition 5.1, and Lemma 5.2 that this note corrects."},{"cited_title":"Douady, J.H","cited_arxiv_id":null,"evidence_quote":"Provides the polynomial-like map theorem and the winding condition that the corrected regions must satisfy."},{"cited_title":"Devaney, Baby Mandelbrot sets adorned with Halos in families of rational maps, Contemporary Mathematics, 396 (2006), 37-50","cited_arxiv_id":null,"evidence_quote":"Establishes the equal-exponent case $n=d$, which the note excludes and uses as a comparison boundary for its correction."}],"review_version":1}