{"id":"af1b7035-c00d-4d89-bd52-5b1d1d5cc240","arxiv_id":"1908.05173","paper_version":4,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For real plane cubics, allowing all polynomial automorphisms instead of only affine maps yields exactly the same classification, with a complete list of normal forms.","lead":"This paper classifies all real cubic curves (degree-three polynomial curves in the plane) up to two notions of symmetry: affine transformations and the broader set of all polynomial automorphisms. It finds that the broader symmetry does not merge any of the affine classes, producing a complete list of 33 representative curves.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 13 rests on an unjustified and generally false claim that Q has real coefficients, so the automorphic pairwise-inequivalence proof for the x^3 family is incomplete.","rationale":"I read the paper as a classification of real cubic polynomials under polynomial automorphisms, with the central claim that every orbit has a representative in Table 13 and that the table has no duplicate automorphic classes. The affine part rests on the four real binary-cubic normal forms; this is a standard true fact, although the paper does not prove it. I checked the reader's Table 14 objection: the row for x^3-y^2-x+J is correct (the value J+2√3/9 is the isolated point at x=-1/√3 and J-2√3/9 is the node at x=1/√3), so that specific complaint does not land. The real weakness is in Proposition 13. The assertion that Q has real coefficients is not implied by ψ=θ∘ϕ∘θ and is false in general, as the triangular example shows. The later coefficient comparison therefore assumes the realness that it needs to prove. Because Proposition 13 is the only proof of pairwise inequivalence for the x^3-family, Theorem 14 is not proven as written. The result may still be correct: Proposition 15 gives a different, plausible Abhyankar-Moh argument for the problematic x^3-y^2+linear cases, and the other cases seem to follow from the invariant sets in Lemma 2. I therefore keep the reader's conditional verdict rather than escalating, but the required correction is to repair Proposition 13 or replace it with Proposition 15.","tokens_in":13469,"tokens_out":32498,"duration_ms":298176,"concrete_test":"Test the disputed step directly: solve the degree-two system (18) without imposing that Q_1 has real coefficients and confirm that (P_1,Q_1)=(-x,iy) is a solution; then determine whether this solution can be extended to an automorphism ψ of the form θ∘ϕ∘θ for a real ϕ. If it cannot, identify the additional constraint coming from real ϕ that excludes it, and check whether Proposition 15's Abhyankar-Moh argument covers all x^3-y^2+linear cases. If no such constraint is found, the c=-1 case of Proposition 13 is a genuine counterexample to the claimed pairwise inequivalence.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Proposition 13 is the only argument that the x^3-family in Table 13 is pairwise inequivalent under ≈. Its treatment of the c=-1 case is invalid. After equation (17) the proof says 'all of the coefficients of Q must be real' for ψ=θ∘ϕ∘θ, where θ(x)=x+i√3/3. This does not follow: for a real automorphism ϕ, ψ(y)=Q_ϕ(x+i√3/3,y), whose coefficients usually lie in R[i√3] rather than R. For example, the real triangular automorphism ϕ(x)=x, ϕ(y)=y+x^2 gives ψ(y)=y+(x+i√3/3)^2, which has non-real coefficients. The subsequent contradiction (equations (19)-(21)) depends on writing Q_1=cx+dy with c,d∈R. If Q_1 is allowed complex, the linear-part system has a solution: P_1=-x, Q_1=iy satisfies (18), and the full map ψ(x)=-x, ψ(y)=iy satisfies (17). The proof also asserts that ψ fixes the origin, but the polynomial x^3-i√3 x^2-y^2 has two singular points, (0,0) and (2i√3/3,0), so the origin is not uniquely forced. Thus Theorem 14's automorphic classification is not established by the written proof.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies two equivalence relations on real plane curves: the affine equivalence ∼ generated by Γ_2(R) and scalings, and the coarser automorphic equivalence ≈ generated by all ring automorphisms of R[x,y] and scalings. It first completes Weinberg's affine classification, producing Table 13 as a complete and pairwise inequivalent list for ∼. It then argues that the same list serves for ≈, i.e. that no two entries of Table 13 become automorphically equivalent, and records a few cubic polynomials with automorphic degree less than three in Table 15. The main theorem is Theorem 14, which asserts exactly this, and the proof is split by the four possible leading homogeneous forms.","tokens_in":13670,"tokens_out":26113,"duration_ms":226535,"significance":"If the main theorem is correct, the paper gives a clean, complete answer to a natural classification problem and shows that the full automorphism group does not coarsen the affine classification for real cubic curves. The affine half of the paper is substantial and appears to be carefully executed, with a full case split and pairwise inequivalence arguments. The automorphic half uses attractive elementary invariants (degree of an automorphic image, sets of critical values with cusp/isole/node types, reducibility sets). However, the proof of Proposition 13, which is load-bearing for the ≈-pairwise-inequivalence of the x^3 family, contains a serious gap; the stated contradiction is not valid. The paper's central claim therefore needs a repaired argument before it is established.","major_comments":[{"comment":"In the proof for the family x^3-y^2+x+J, the assertion after (17) that 'all of the coefficients of Q must be real' is false. For ψ=θ∘ϕ∘θ with θ(x)=x+i√3/3, the coefficient field of Q=ψ(y) is R[i√3] in general, not R; for instance, with the real automorphism ϕ=⟨x, y+x^2⟩, Q=y+(x+i√3/3)^2 has the non-real coefficient 2i√3/3. Consequently the subsequent decomposition Q_1=cx+dy with c,d∈R in equations (19)-(21) is not justified, and no contradiction is obtained. In fact the C-automorphism ψ=⟨-x, iy⟩ fixes C and satisfies (17) as written, so (17) cannot be used to derive an impossibility. The proof of pairwise inequivalence for this family is therefore incomplete.","section":"§2, Proposition 13 (set (12))"},{"comment":"The proof also claims that θ∘ϕ∘θ must fix the origin because it sends the singular point of x^3-i√3 x^2-y^2 to the singular point of -x^3-i√3 x^2+y^2. This is not forced: the source polynomial has two singular points, (0,0) and (2i√3/3,0), and the target polynomial also has two singular points, so an automorphism may interchange them. Hence P_0=Q_0=0 is not justified. This is a second independent reason the written argument for the x^3-y^2+x+J family does not go through.","section":"§2, Proposition 13 (set (12))"},{"comment":"Every proposition in the affine classification begins with the assumption that the homogeneous degree-three part can be brought by a linear change of coordinates to one of x^3+xy^2, x^3-xy^2, x^2y, or x^3. The paper states this only as 'As in [10], we will assume' without proof or a precise citation. The statement is true (it follows from the classification of real binary cubics by discriminant and root multiplicities), but because Theorem 9 and hence Theorem 14 depend on this case split, a proof or an exact statement from [10] should be supplied.","section":"§1, initial case split"}],"minor_comments":[{"comment":"In the rows x^3-xy+1 and x^3-xy, the node(f) column is listed as empty. For r=1 and r=0 respectively, f-r equals x^3-xy=x(x^2-y), whose components x=0 and y=x^2 meet transversely at the origin, so the node sets should be {1} and {0}. The logical argument is unaffected because red(f) is used for these rows, but the table should be corrected.","section":"Table 14"},{"comment":"The use of the Abhyankar-Moh theorem is not justified as written: P(0,y) and Q(0,y) need not generate C[y], which is required for the degree-divisibility conclusion. For example, the automorphism (x+y^2,y) restricts to (y^2,y). Since Proposition 15 is not needed for Theorem 14, this is a non-blocking issue, but the proposition should be repaired or removed.","section":"§2, Proposition 15"},{"comment":"The displayed identity (17) does not appear to follow from the preceding definitions. With θ(x)=x+i√3/3, the composition θ^{-1}(-x^3+y^2-x+2i√3/9) equals -x^3+i√3 x^2+y^2+4i√3/9, not -x^3-i√3 x^2+y^2. The sign and constant terms in (17) should be checked.","section":"§2, Proposition 13"},{"comment":"Minor typographical issues: 'repectively' should be 'respectively' in Proposition 15, and 'the the xy-coefficient' in Proposition 8 should be 'the xy-coefficient'.","section":"§2, Proposition 15; §1, Proposition 8"}],"recommendation":"major_revision","confidential_remarks":"The paper is concise and the affine part is solid; the automorphic proof has a gap in Proposition 13 that is likely repairable. I would recommend major revision with an eye toward a correct proof for the x^3-y^2+x+J family. The canonical-form assumption should be proved or explicitly referenced, and the Abhyankar-Moh alternative in Proposition 15 should not be used until the generation condition is justified."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe short version: the affine half of this paper is solid and genuinely useful, and the automorphic result is probably true, but the written proof of the automorphic part has a gap that matters. Note first that the reader's Table 14 complaint is a false alarm—the table correctly lists the node at J−2√3/9 and the isolated point at J+2√3/9 for x^3−y^2−x+J. The real problem is Proposition 13.\n\nWhat's new and good: the affine classification re-derives and completes Weinberg's list with explicit normal forms and a clear case analysis. The observation that the full automorphism group Aut(R[x,y]) collapses only the few low-degree cases in Table 15 (e.g., x^3−y maps to x) is interesting and worth recording. The invariant sets red, sing, cusp, isol, node and Lemma 2 are a sensible tool, and the affine pairwise-inequivalence proofs in Propositions 5–8 are terse but mostly checkable.\n\nThe soft spot is load-bearing. In Proposition 13, the c=−1 case for the x^3−y^2+x+J family, the proof conjugates by θ(x)=x+i√3/3 and then asserts 'all of the coefficients of Q must be real' for ψ=θ∘ϕ∘θ. That is false for a real ϕ. For instance, if ϕ(x)=x and ϕ(y)=y+x^2, then ψ(y)=y+(x+i√3/3)^2, which has complex coefficients. The subsequent contradiction in equations (19)–(21) depends on writing Q_1=cx+dy with real c,d, but the linear system has complex solutions—ψ(x)=−x, ψ(y)=iy satisfies (17). The proof also says ψ 'must fix the origin' because it sends the singular point to the singular point, but x^3−i√3x^2−y^2 has two singular points, (0,0) and (2i√3/3,0). So the automorphic inequivalence of the x^3 family is not established by the written argument.\n\nMinor issues: the unproved assumption that the homogeneous cubic part is one of four canonical forms is standard and true over R, but it should be stated as a lemma or cited precisely. Several 'by inspection' steps could use more detail.\n\nBottom line: a plausible result, worth publishing once Proposition 13 is repaired—possibly using the Abhyankar–Moh argument the author already sketches as Proposition 15. As it stands, the paper deserves serious peer review, but it should not be accepted without revision.","headline":"Solid affine classification, plausible automorphic result, but Proposition 13's key real-coefficient claim is false, so Theorem 14 is not yet proven.","tokens_in":14256,"tokens_out":12113,"would_cite":false,"duration_ms":102264,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14N99","14H10"],"pacs":[],"model":"deepseek-v4-flash","headline":"Every real cubic plane curve is equivalent, under all polynomial automorphisms of the affine plane, to exactly one normal form in a single table, namely the same table that classifies cubics up to affine transformations.","keywords":["real cubic curves","polynomial automorphisms","affine classification","normal forms","plane curves","ring automorphisms","automorphism degree"],"falsifier":"Work out the orbit classification of real homogeneous cubic forms in two variables under real linear changes and nonzero scalars. The four forms $x^3 + xy^2$, $x^3 - xy^2$, $x^2y$, $x^3$ correspond to the possible real root patterns (one real root, three real roots, a double root, a triple root); if this list is complete, the paper's opening assumption holds, and if it is not, the normal-form table omits classes. The paper states the assumption without proof, so this direct orbit computation is the decisive check.","tokens_in":13187,"feed_emoji":"📐","tokens_out":12209,"duration_ms":118300,"temperature":0.7,"pith_summary":"The paper aims to complete the classification of real plane cubic curves under the action of all ring automorphisms of $\\mathbb{R}[x,y]$, up to nonzero scalar. It proves that every degree-three polynomial is equivalent to exactly one polynomial in Table 13, and that the polynomials in Table 13 are pairwise inequivalent. Because Table 13 is the same list used for the affine classification, the automorphic classification and the affine classification coincide for real cubics. Along the way it finishes the affine classification begun in [10] and identifies the three cubics that can be mapped by polynomial automorphisms to polynomials of lower degree.","feed_headline":"One table holds every real cubic curve class","feed_subtitle":"Allowing arbitrary polynomial automorphisms adds no new classes beyond the affine classification.","key_machinery":"The load-bearing object is the orbit degree $\\operatorname{AutDeg}(f) = \\min\\{\\deg(g) : g \\approx f\\}$, together with the invariant sets $\\operatorname{cusp}(f)$, $\\operatorname{isol}(f)$, $\\operatorname{node}(f)$, $\\operatorname{red}(f)$, and $\\operatorname{sing}(f)$ that record for which constants $r$ the curve $f - r = 0$ acquires a cusp, an isolated point, a node, reducibility, or a singularity. These invariants are carried through polynomial automorphisms by Lemma 2, so they separate table entries that degree arguments cannot. The proof strategy is to normalize the homogeneous cubic part to one of four canonical forms, use the affine classification to reach Table 13, and then rule out non-affine automorphisms by comparing degrees of the image polynomials $p, q$ and by using the invariant sets. A separate argument for the remaining $x^3 - y^2 + \\lambda(f)$ case applies the epimorphism theorem for embeddings of the line in the plane.","core_discovery":"The paper's central claim is Theorem 14: with respect to $\\approx$, the action of all ring automorphisms of $\\mathbb{R}[x,y]$ up to scalar, every cubic polynomial is equivalent to one entry of Table 13, and the entries of Table 13 are pairwise inequivalent. Because affine equivalence $\\sim$ is a special case of $\\approx$, this says the automorphic classification of real cubic curves coincides with the affine classification. The proof first completes the affine classification, then shows that any automorphic equivalence between two table entries forces the automorphism to be affine. The exceptional cases are $x^3 - y \\approx x$, $x^3 - xy \\approx xy$, and $x^3 - xy + 1 \\approx xy + 1$; these are the only cubics for which a polynomial automorphism can lower the degree below three.","pith_inferences":["If the same invariant-plus-degree strategy were applied to real quartic plane curves, the automorphic and affine classifications would likely diverge: the paper's degree-lowering examples show that polynomial automorphisms can change degree, and the invariant sets used here become harder to compute as the degree rises.","The unproved four-form assumption is exactly the classification of real binary cubic forms under linear change; an explicit proof of that orbit list would remove the one gap in the paper's foundation.","The explicit automorphisms behind the degree-lowering examples provide ready-made test cases for algorithms that compute automorphism orbits or detect equivalences between plane curves.","One could test the classification computationally by sampling cubics, running the case-table normalization, and checking that the output lies in Table 13; a failure would pinpoint a gap, while success over a broad sample would corroborate Theorem 14."],"forward_implications":["Table 13 serves simultaneously as the affine and the automorphic normal-form list for real cubic curves.","The only real cubics whose automorphism orbit contains a polynomial of degree less than three are $x^3 - y$, $x^3 - xy$, and $x^3 - xy + 1$.","The invariants $\\operatorname{cusp}$, $\\operatorname{isol}$, $\\operatorname{node}$, $\\operatorname{red}$, and $\\operatorname{sing}$ are invariants of the full automorphism group, not just the affine group, so they can be used to distinguish plane curves of other degrees.","Because the normal forms are parameterized by inequalities, deciding whether two given cubics are equivalent reduces to putting each in Table 13 and comparing parameters."],"supporting_citations":[{"why":"Supplies the affine classification that the paper completes and the four canonical homogeneous forms used throughout.","marker":"[10]"},{"why":"Definition 4, the invariant sets cusp, isol, node, red, and sing, is adapted from this source and is used in Proposition 13.","marker":"[1]"},{"why":"Supplies the line-embedding epimorphism theorem used in Proposition 15 to rule out non-linear automorphisms of cubics of the form $x^3 - y^2 + \\lambda(f)$.","marker":"[2]"}],"fun_headline_variants":["Polynomial automorphisms add no new real cubic classes","Automorphic and affine classifications coincide for real cubics","Cubic classification resists all polynomial automorphisms","Polynomial maps keep real cubic curve classes unchanged"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper assumes without proof that every homogeneous cubic in two real variables can be moved by a real linear change of coordinates to one of four forms, $x^3 + xy^2$, $x^3 - xy^2$, $x^2y$, or $x^3$; if some real binary cubic lay outside these four orbits, Table 13 would miss entire equivalence classes.","fun_headline_variants_meta":{"raw":{"variants":["Polynomial automorphisms add no new real cubic classes","Automorphic and affine classifications coincide for real cubics","Cubic classification resists all polynomial automorphisms","Polynomial maps keep real cubic curve classes unchanged"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000649,"raw_usage":{"total_tokens":2864,"prompt_tokens":718,"completion_tokens":2146,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":334,"completion_tokens_details":{"reasoning_tokens":2083}},"tokens_in":334,"tokens_out":2146,"duration_ms":14741,"temperature":1.0,"reasoning_tokens":2083,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:24:53.703930+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Work out the orbit classification of real homogeneous cubic forms in two variables under real linear changes and nonzero scalars. The four forms $x^3 + xy^2$, $x^3 - xy^2$, $x^2y$, $x^3$ correspond to the possible real root patterns (one real root, three real roots, a double root, a triple root); if this list is complete, the paper's opening assumption holds, and if it is not, the normal-form table omits classes. The paper states the assumption without proof, so this direct orbit computation is the decisive check.","supporting_citations":[{"cited_title":"W einberg, The affine classification of cubic curves , The Rocky Mountain Journal of Mathematics","cited_arxiv_id":null,"evidence_quote":"Supplies the affine classification that the paper completes and the four canonical homogeneous forms used throughout."},{"cited_title":"Abhyankar, W.J","cited_arxiv_id":null,"evidence_quote":"Definition 4, the invariant sets cusp, isol, node, red, and sing, is adapted from this source and is used in Proposition 13."},{"cited_title":"Abhyankar and T.T","cited_arxiv_id":null,"evidence_quote":"Supplies the line-embedding epimorphism theorem used in Proposition 15 to rule out non-linear automorphisms of cubics of the form $x^3 - y^2 + \\lambda(f)$."}],"review_version":1}