{"id":"f84c2c78-15a7-4026-ac7e-f30a4a927757","arxiv_id":"1908.04117","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper conjectures that a certain pencil of algebraic cycles on degree-8 Fermat surfaces yields infinitely many distinct special Noether-Lefschetz components, contradicting Harris's finiteness conjecture, and rigorously proves infinite general components for degrees 4 through 9.","lead":"These math results study special surfaces in three-dimensional space that have extra algebraic curves, and propose a specific infinite family that would disprove a 1980s conjecture by J. Harris. The paper gives computational evidence and proves related infinite families for Fermat surfaces of degree up to 9.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The exceptional special-pencil claim rests on an unproved finite-to-all-r extrapolation: 5-jet smoothness for |r1|,|r2|≤10 does not establish analytic smoothness for all but finitely many r, and 6-smoothness is explicitly left open.","rationale":"The reader's CONDITIONAL verdict is appropriate. The rigorous components (Proposition 1, Theorem 3, the formula in §5) are sound as far as the text shows, and the exceptional-case claim is explicitly phrased as evidence/conjecture rather than as a proof. My review identifies the same load-bearing weakness as the reader: the finite range (7) and the unverified 6-jet are the point where the special-pencil conclusion outruns the computation. A polynomial-root computation for transversality would turn the finite check into an exact statement about all rational r, removing the main extrapolation. Agreement is \"agree\" because this matches the reader's weakest_assumption. The verdict stays CONDITIONAL: the paper should be accepted only with access to the external Singular code/data and with the 6-smoothness gap openly stated; no change from the reader's verdict is warranted.","tokens_in":11353,"tokens_out":16636,"duration_ms":189630,"concrete_test":"Using the period formula (13), form the period matrix P(r)=[p_{i+j}([C1]+r[C2])] for the d=8 exceptional case (6). The transversality of T0V_{[C1]+r[C2]} to the 32-dimensional deformation space (9) is equivalent to the condition that the submatrix with row set I* from §4 has the same rank as P(r). Compute the maximal minors of this submatrix as polynomials in r, take their gcd over Q, and factor it; list all rational roots. If the root set is contained in {0,±1}, the finite-to-all-r transversality extrapolation of Remark 3 is settled for every rational r. If any new root appears, that r is an unaccounted exception and the stated evidence for the special pencil must be revised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central object of the paper is the claim that the family (6) is a special pencil, i.e., that V_{[C1]+r[C2]} is smooth as an analytic scheme for all but finitely many r∈Q (Definition 1). The strongest verified statement, Theorem 2, is only that V^5_{[C1]+r[C2]} is the 5-jet of a smooth variety, and only for r in the finite range (7), inside the 32-dimensional deformation space (9). Section 6 and Table 1 (row d=8) explicitly record that 6-smoothness was not verified. No theorem is given showing that 5-jet smoothness implies smoothness of the analytic germ, and no exhaustive computation covers all rational r. The extrapolation recorded in Remark 3 (\"for d=8 we only need to exclude r1, |r2|=0,1\") is itself a finite check over (7). Thus a single rational r outside the tested range with either non-transversality to the smaller deformation space or failure of 6-smoothness would be an unaccounted exception; if there are infinitely many, the special-pencil conclusion fails, while finitely many would leave the evidence incomplete. The paper honestly labels the main assertion conjectural, so this is a gap in evidence rather than a contradiction; it is exactly the condition on which the central claim depends.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies Noether-Lefschetz loci on Fermat surfaces of degrees 4 through 11, concentrating on the pencils of classes [C1] + r[C2], r ∈ Q, where C1 is a line and C2 is a (3,3) complete intersection curve on a smooth degree-8 surface with C1 ∩ C2 = ∅. The main object is the conjecture that, for the exceptional choice (6), these loci form a special pencil: for all but finitely many r, V_[C1]+r[C2] is a smooth, reduced, codimension-31 special component, and distinct members meet in a codimension-32 subvariety, yielding a counterexample to Harris's conjecture. Rigorous results include Proposition 1, a criterion for general pencils; Theorem 3, which produces many general pencils for d ≤ 9 and none for d = 10, 11; Theorem 1, which detects singularities of many loci for r in the finite range (7); and Theorem 2, which shows that in the exceptional d = 8 case the 5th infinitesimal locus in a chosen 32-dimensional deformation space is the 5-jet of a smooth variety. The paper explicitly labels the special-pencil assertion as conjectural, and the evidence for it is finite and computational.","tokens_in":11677,"tokens_out":11557,"duration_ms":129857,"significance":"If the conjectural construction could be completed, it would give infinitely many reduced special components of NL_8 passing through the Fermat point, refuting Harris's conjecture in degree 8; this makes the paper potentially quite significant. The rigorously proved parts are also valuable: Proposition 1 is a clean general-pencil criterion, and Theorem 3 provides many explicit general pencils through the Fermat point, including an infinite supply of general components for degrees ≤ 9. A clear strength is the computational transparency: code1, code2, code3 and the Singular library foliations.lib are referenced and linked, and the paper explains how the data in Table 1 were generated. On the other hand, the central special-pencil claim is not a theorem; it depends on finite computations, on an unverified 6-jet, and on an unproved extrapolation from a finite range of r to all but finitely many r ∈ Q. The honest labeling of the main assertion as a conjecture is appropriate, but the published contribution is a body of computational evidence rather than a proof of the counterexample.","major_comments":[{"comment":"Theorem 2 is proved only for the restriction of the infinitesimal locus to the 32-dimensional deformation space (9), not for the full parameter space T from (14). The transversality discussion in §4 and the implication in §6 are one-way: the paper uses failure of smoothness of the restricted locus to conclude non-smoothness in T. No converse statement is proved. Therefore Theorem 2 by itself does not establish even 5-smoothness, let alone analytic smoothness, of V_[C1]+r[C2] in the full parameter space that appears in Definition 1.","section":"§4, §6, Theorem 2"},{"comment":"Even if full 5-smoothness were available, Definition 1 requires the analytic germ to be smooth, and the paper does not prove that 5-smoothness implies analytic smoothness. For the decisive exceptional case (6), Table 1 leaves the N=6 and N≥7 columns as '?', and §6 states explicitly that the author was not able to verify 6-smoothness. This is a load-bearing gap: the missing next jet is exactly the condition on which the special-pencil claim and the potential counterexample to Harris's conjecture depend.","section":"Definition 1, §6, Table 1 (row d=8)"},{"comment":"The exclusion of non-transversal pairs is checked only on the finite range (7), with the statement in Remark 3 that for d=8 'we only need to exclude r1, |r2| = 0,1' being an observation about that finite range. Lower semicontinuity of the rank gives that the codimension of the tangent space is maximal away from a finite set in P^1, but it does not identify that finite set and it says nothing about N-smoothness for N ≥ 2. Without a proven bound, or a certificate for an exhaustive computation, the conclusion 'for all except a finite number of r' in Definition 1 is not justified from the stated evidence.","section":"§6, (7), Remark 3"},{"comment":"The proofs of Theorems 1 and 3 depend on code1, code2, code3 and the Singular library foliations.lib, with data linked on the author's webpage. The availability of code and data is commendable, but the manuscript does not specify versions of Singular and foliations.lib, and it does not provide a verification log or certificate. Since a bug or configuration issue in the external library cannot be distinguished from a genuine mathematical fact on the basis of the printed text, an independent implementation or a machine-checked computation log would substantially strengthen the computational theorems.","section":"§8"}],"minor_comments":[{"comment":"The quantifier in Theorem 1 is implicit; the conclusion should state explicitly that it holds for all pairs (r1, r2) satisfying both (7) and (8).","section":"Theorem 1"},{"comment":"The entries '1220+113' in the d=8 row and '1314+19' in the d=9 row should be explained; it is not clear whether these denote a partition of the cases or a typographical convention.","section":"Table 1"},{"comment":"In §8, the example command 'DeformSapce' is a typo for 'DeformSpace', and in Remark 6 the sentence 'V^2_[C1]+r[C2] the 2-jet of a smooth variety' is missing a verb.","section":"§8 and Remark 6"},{"comment":"The references [Mov19] and [MV19] should include complete bibliographic information, including the version or expected publication details, since the paper relies on them for several foundational statements.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"This is a borderline case. The author is honest that the central special-pencil claim is conjectural, and the rigorous results and computational transparency are genuine strengths. However, the published text should make unmistakably clear that the main contribution is a body of computational evidence for a conjecture, not a proof. Given the reliance on the author's own software and unpublished notes, the editor may wish to seek an independent computational check of the tables before considering acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThis paper deserves a serious read, but the headline claim is a conjecture, not a theorem. The special pencil concept is new, and the degree-8 configuration (a line plus a (3,3) complete intersection, disjoint, on the Fermat surface) is a clever candidate to break Harris's conjecture. What is actually proven is more modest but still valuable: Proposition 1 and Theorem 3 establish an infinite number of general components through the Fermat point for d=4..9, and their absence for d=10,11. Those are rigorous, and the data is made available through hyperlinks and Singular code.\n\nThe paper is honest about what is open. Section 6 and Table 1 explicitly record that 6-smoothness was not verified for the exceptional case, and the main assertion is labeled as gathered evidence. That honesty is a real point in its favor.\n\nThe soft spot is the leap from 5-jet smoothness for |r1|,|r2|≤10 to analytic smoothness for all but finitely many rational r. There is no theorem bridging N-jet smoothness to smoothness of the analytic germ, and no exhaustive check over all r. The claim in Remark 3 that only r1,|r2|=0,1 need exclusion for d=8 is itself an extrapolation from a finite computation. A single rational r outside the tested range with non-transversality or a smoothness failure would sink the special-pencil conclusion; if there are infinitely many, the conjecture fails. The stress-test note is right about this.\n\nA second barrier is reproducibility. The computational theorems depend on code and data outside the arXiv version. For results that are load-bearing in this context, that makes independent verification hard. The author should either provide the key scripts in the submission or a detailed log of the d=8 runs.\n\nCircularity concerns are minor. The deformation space (9) is chosen specifically to make V[C1]∩V[C2]={0}, which is a legitimate construction but limits what can be inferred about the full parameter space. The paper does not pretend otherwise.\n\nThis paper is for Hodge theorists and anyone following Harris's conjecture. I would send it to peer review. The rigorous components are new and worth having on record, and the conjectural part can be sharpened under referee pressure: ask for the 6-smoothness check or an explicit finite set of candidate exceptions, and for the code to be bundled. A serious referee can handle that.","headline":"Genuinely new rigorous results on general Noether-Lefschetz pencils, but the special-pencil counterexample to Harris is a well-labeled conjecture resting on finite computational evidence.","tokens_in":12150,"tokens_out":3849,"would_cite":true,"duration_ms":37886,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14C30","14D07","14J70"],"pacs":[],"model":"deepseek-v4-flash","headline":"On the Fermat surface of degree 8, a rational pencil of algebraic cycles is conjectured to produce infinitely many special Noether–Lefschetz components, contradicting a 1980s finiteness conjecture.","keywords":["Noether-Lefschetz locus","special components","finiteness conjecture","Hodge cycles","Fermat surface","infinitesimal variation of Hodge structure","period integrals","algebraic cycles"],"falsifier":"Compute the sixth-order infinitesimal Noether–Lefschetz locus $V^6_{[C_1]+r[C_2]}$ in the 32-dimensional deformation space for the exceptional case and a rational $r$ with $r_1>10$ or $|r_2|>10$; if it is not the 6-jet of a smooth variety, or if the tangent space loses transversality to the deformation space for some such $r$, the conjectured special pencil fails. Alternatively, a single $r$ outside the checked range for which $V_{[C_1]+r[C_2]}$ has codimension different from 31 would disprove the pencil claim.","tokens_in":11116,"feed_emoji":"📐","tokens_out":8593,"duration_ms":79491,"temperature":0.7,"pith_summary":"The paper studies a one-parameter family of algebraic cycles on the Fermat surface of degree 8, namely $[C_1]+r[C_2]$ with $r\\in\\mathbb{Q}$, where $C_1$ is a line and $C_2$ a complete intersection curve of type $(3,3)$ disjoint from it. It argues, on the basis of extensive exact computation, that for all but finitely many $r$ these Noether–Lefschetz loci are distinct $31$-codimensional subvarieties meeting pairwise in a $32$-codimensional subvariety of the parameter space. Since the maximal possible codimension for a Noether–Lefschetz component here is $35$, such a pencil would describe infinitely many reduced special components through a single point, contradicting the conjecture that special components are finite. The rigorous theorems establish surrounding structure: singularity of the loci for most tested $r$ in degrees $4$ through $8$, a fifth-order smoothness statement in the exceptional degree-8 case, and the existence of many general pencils for degrees $4$ through $9$ and none in degrees $10,11$.","feed_headline":"A special pencil in degree 8 would refute a finiteness conjecture","feed_subtitle":"A rational pencil of curves on the Fermat octic yields distinct 31-codimensional loci for all but finitely many r","key_machinery":"Two computational tools carry the argument. First, the tangent space of a Noether–Lefschetz locus at the Fermat point is identified with the kernel of the period matrix $[p_{i+j}(\\delta)]$ coming from infinitesimal variation of Hodge structure, so codimensions of $V_{[C_1]+r[C_2]}$, of its intersections, and inclusions among its tangent spaces can be computed exactly from periods of the two curves. Second, for the exceptional degree-8 case the paper derives an explicit Taylor series (equation (23)) for the integrals of differential forms over the monodromy of a rational curve, expressed through Pochhammer symbols and power sums of roots of unity; this formula lets the computation check, order by order, whether $V^N_{[C_1]+r[C_2]}$ is the $N$-jet of a smooth variety. The chosen 32-dimensional deformation space is constructed so that $T_0V_{[C_1]}\\cap T_0V_{[C_2]}=\\{0\\}$, making transversality of the pencil visible inside that subspace.","core_discovery":"The paper's central object is the Noether–Lefschetz locus $V_{[C_1]+r[C_2]}$ attached to the rational cohomology class of the algebraic cycle $C_1+rC_2$ inside a smooth degree-8 surface, with $C_1$ a line and $C_2$ a $(3,3)$ complete intersection curve disjoint from it. The main thesis is that for the Fermat surface of degree 8 and this choice of $C_1,C_2$, and for all but finitely many rational numbers $r$, these loci are distinct, reduced, $31$-codimensional subvarieties of the parameter space of degree-8 surfaces, pairwise intersecting in a $32$-codimensional subvariety. This would be a special pencil in the paper's terminology, and because the maximal codimension of a Noether–Lefschetz component here is $35$, it would yield infinitely many strictly special reduced components through a single point, contradicting the 1980s conjecture that special components are finite. The paper does not prove the full statement; it proves Theorem 2, that in the exceptional case the fifth-order infinitesimal locus $V^5_{[C_1]+r[C_2]}$ is the 5-jet of a smooth variety in a carefully chosen 32-dimensional deformation space, and Theorem 1, that in many other degree $4$–$8$ cases the loci are singular for most $r$ in a finite tested range.","pith_inferences":["Beyond the paper: the same line-plus-complete-intersection construction could be tried on Fermat hypersurfaces of higher degree and even dimension; Section 7 already indicates where general Hodge loci live, and the special-pencil mechanism should transfer whenever two Hodge cycles have independent tangent spaces.","Beyond the paper: if the conjectured pencil is real, the uniform 32-codimensional intersection would impose a strong relation among the period matrices of $C_1$ and $C_2$, possibly making the whole family detectable from a single higher-order jet.","Beyond the paper: the finite-range computation could be converted into a proof if the relevant matrix rank is shown to be a polynomial in $r$ with no zeros outside a finite set; Theorem 2 would then become the full special-pencil statement."],"forward_implications":["If the exceptional degree-8 case is a special pencil, the Noether–Lefschetz locus of degree-8 surfaces has infinitely many reduced special components passing through the Fermat point, falsifying the 1980s finiteness conjecture.","Theorem 1 implies that for degrees 4 through 7, and for many degree-8 families, every rational cycle in the tested range with $3\\le r_1$ or $3\\le |r_2|$ gives a singular Noether–Lefschetz locus, so no special pencil arises from those families.","Theorem 3 gives infinitely many general (maximal-codimension, smooth, reduced) Noether–Lefschetz components through the Fermat point for degrees 4 through 9, while for degrees 10 and 11 none of the tested pencils are general.","The explicit period formula provides a finite, checkable criterion for $N$-smoothness of the infinitesimal Noether–Lefschetz locus, so the central conjecture is algorithmically testable at any fixed jet order."],"supporting_citations":[{"why":"Supplies the infinitesimal variation of Hodge structure framework from which the tangent-space description of the Noether–Lefschetz locus is derived.","marker":"[CGGH83]"},{"why":"Provides a prior counterexample to the finiteness conjecture for large degree, which this paper's special-pencil conjecture would extend to degree 8.","marker":"[Voi91]"},{"why":"Establishes that the number of reduced special components is finite for degrees 6 and 7, the contrast case for the degree-8 claim.","marker":"[Voi90]"},{"why":"Gives the period formula for complete intersection curves on the Fermat surface used to compute the codimensions in Table 1.","marker":"[MV19]"},{"why":"Provides the Taylor series for periods over monodromy and the finite criterion for $N$-smoothness that the computer checks.","marker":"[Mov19]"},{"why":"Proves density of general components, fixing the baseline notion of general against which special pencils are measured.","marker":"[CHM88]"}],"fun_headline_variants":["Special pencil on degree 8 would refute Harris finite-loci conjecture","Octic pencil: distinct 31-codim Noether-Lefschetz loci for all but finitely many r","New evidence against Harris conjecture from Fermat octic pencil","Fermat octic pencil: special Noether-Lefschetz loci for all but finitely many r","Potential counterexample to Harris conjecture via degree-8 pencil"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The claim that the exceptional degree-8 case gives a special pencil rests on checking transversality and $N$-jet smoothness only for rational parameters $r=r_2/r_1$ with $1\\le r_1\\le10$ and $0\\le|r_2|\\le10$, then extrapolating to all but finitely many rational $r$.","fun_headline_variants_meta":{"raw":{"variants":["Special pencil on degree 8 would refute Harris finite-loci conjecture","Octic pencil: distinct 31-codim Noether-Lefschetz loci for all but finitely many r","New evidence against Harris conjecture from Fermat octic pencil","Fermat octic pencil: special Noether-Lefschetz loci for all but finitely many r","Potential counterexample to Harris conjecture via degree-8 pencil"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001004,"raw_usage":{"total_tokens":4303,"prompt_tokens":1055,"completion_tokens":3248,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":671,"completion_tokens_details":{"reasoning_tokens":3141}},"tokens_in":671,"tokens_out":3248,"duration_ms":22096,"temperature":1.0,"reasoning_tokens":3141,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:51:07.785905+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the sixth-order infinitesimal Noether–Lefschetz locus $V^6_{[C_1]+r[C_2]}$ in the 32-dimensional deformation space for the exceptional case and a rational $r$ with $r_1>10$ or $|r_2|>10$; if it is not the 6-jet of a smooth variety, or if the tangent space loses transversality to the deformation space for some such $r$, the conjectured special pencil fails. Alternatively, a single $r$ outside the checked range for which $V_{[C_1]+r[C_2]}$ has codimension different from 31 would disprove the pencil claim.","supporting_citations":[],"review_version":1}