{"id":"2a98f367-1a9a-4df3-801b-ff3bdb2d680b","arxiv_id":"2607.22233","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A condition on Lefschetz primitive components of algebraic odd cohomology implies incidence equivalence equals Abel-Jacobi equivalence, repairing a gap in Murre's codimension-two proof.","lead":"This survey explains when a geometric way of comparing cycles—incidence equivalence—matches a transcendental one called Abel-Jacobi equivalence. It also adds a new Hodge-theoretic condition that would repair a gap in a classical codimension-two proof.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Equation (14), the step on which Corollary 7.8 hinges, is not established: the definition of H_alg^{2i-1} as N^i H^{2i-1} is inconsistent with the paper's own coniveau convention, and the star/Weil preservation of algebraic classes is asserted without checking coniveau levels.","rationale":"The reader correctly targets equation (14) as the load-bearing step. My pass strengthens that objection: there is a prior indexing problem. Equation (12) defines the odd algebraic cohomology H_alg^{2i-1} as N^i H^{2i-1}, but under the coniveau filtration defined in §6 this subspace is zero, because a support of codimension at least i has real codimension at least 2i, exceeding the degree 2i-1. Hence the proof cannot be using the objects it appears to define; some reindexing or a different definition is forced. After the natural correction to N^{i-1} H^{2i-1}, the star preservation claim is no longer formal: L preserves coniveau level rather than raising it, so additional input is needed to show *v lands in the required algebraic subspace of the dual degree. This is exactly the kind of missing derivation that a conditional acceptance should require. I do not see evidence of circularity or fabrication, and the strategy—prove non-degeneracy of the odd cup product via Hodge-star positivity—is plausible and worth checking. The reader's CONDITIONAL verdict remains the right one; this note does not change it, but it sharpens the requested fix. Credit is due for a clearly motivated conditional criterion and for citing the corrections to Müller-Stach's theorem.","tokens_in":16865,"tokens_out":26499,"duration_ms":242565,"concrete_test":"Check (14) on a concrete fourfold X with h^{2,1}(X) ≠ 0. Take v = ι_*η, where ι: D → X is a smooth divisor and η ∈ H^1(D, Q), so v ∈ N^1 H^3(X). Compute *v using the Weil formula quoted in §7 and determine the smallest coniveau level of the resulting class in H^5(X). If *v ∉ N^2 H^5, then (14) fails for i=2. Independently, verify whether N^i H^{2i-1}(X) is nonzero for any smooth projective X; if it is always zero, equation (12) must be corrected before any of the Λ(X,i)-statements has content.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Corollary 7.8 depends entirely on (14): for v in H_alg^{2i-1}(X), one needs *v in H_alg^{2d-2i+1}(X) and v·(*v)>0. Positivity is standard, but algebraic preservation is not derived. The one-line appeal to Weil's p.76 formula silently assumes (a) the Weil operator C preserves the rational algebraic subspace, and (b) each L^{d-2i+1+a}C(v_{2i-1-2a}) term lies in the algebraic subspace of the dual odd degree. Neither follows from Condition (Λ,X,i) alone under the definitions as written. In fact (12) defines H_alg^{2i-1}(X) := N^i H^{2i-1}(X). With the coniveau filtration introduced in §6, N^i H^{2i-1}(X) = 0: an algebraic subset of codimension at least i has real codimension at least 2i, which exceeds the degree 2i-1, so the relevant local cohomology vanishes. Thus the definition must be reindexed, e.g. to N^{i-1} H^{2i-1}, or replaced by the maximal level-≤1 Hodge substructure. Once reindexed, the star formula no longer obviously lands in the correct coniveau level: L does not increase coniveau by the projection formula, while the target H_alg^{2d-2i+1} has a different index. Example: in a 4-fold with i=2, a Gysin class from a divisor lies in N^1 H^3, but (14) would require its star to lie in N^2 H^5; the Weil formula only gives a class supported on the same divisor. So (14), and with it the non-degeneracy argument in Corollary 7.8, is unsupported as written.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper surveys Griffiths' incidence equivalence for cycles algebraically equivalent to zero on smooth complex projective varieties, and its relation to Abel-Jacobi equivalence, denoted GC(X,i). It reviews Chow groups, correspondences, standard conjectures, intermediate Jacobians, and Müller-Stach's biextension approach. The main new assertion is Corollary 7.8: a condition (Λ,X,i) on the Lefschetz components of odd-degree algebraic cohomology classes implies GC(X,i); the proof rests on a Hodge-star computation recorded as equation (14). The paper also claims consequences for codimension two cycles, for abelian varieties, and for odd-dimensional varieties with only one middle odd Betti number.","tokens_in":17327,"tokens_out":21296,"duration_ms":207745,"significance":"If the new criterion is correct, it gives a clean structural explanation of Murre's codimension-two theorem and a useful bridge between incidence equivalence, the generalized Hodge conjecture, and the standard conjectures. The survey is well organized and collects relevant older literature, which is valuable. However, the proof of the central new result is currently a sketch with a definitional inconsistency and an unproved Hodge-star step; as written, Corollary 7.8 is not established. The underlying strategy is plausible and may be repairable, but the manuscript in its present form needs substantial corrections before the main claim can be accepted.","major_comments":[{"comment":"The definition H_alg^{2i-1}:=N^i H^{2i-1} is incompatible with the coniveau convention stated earlier in §6. For Z⊂X of codimension at least i, the real codimension is at least 2i, so the local cohomology H^m_Z(X) vanishes for m<2i; in particular H^{2i-1}_Z(X)=0. Hence H_alg^{2i-1}=0 for all i, making Condition (Λ,X,i) vacuous and Lemma 6.3 false (it would imply J^i_alg=0). The definition must be reindexed, e.g. H_alg^{2i-1}=N^{i-1}H^{2i-1}, and then Corollary 6.2, Criterion 7.3, and Condition (Λ,X,i) must be rechecked under the corrected convention.","section":"§6, Eq. (12)"},{"comment":"The assertion 'if v∈H_alg^{2i-1} then *v∈H_alg^{2d-2i+1}' is not derived. Weil's p.76 formula writes *v as a combination of L^{d-2i+1+r} C(v_{2i-1-2r}); the proof does not show that the Weil operator C preserves the relevant algebraic subspace, nor that L sends the corrected coniveau level into the required one. With the natural correction H_alg^{2i-1}=N^{i-1}, a Gysin class supported on codimension i-1 has star supported on the same subvariety, so *v∈N^{i-1}H^{2d-2i+1}, while H_alg^{2d-2i+1}=N^{d-i}H^{2d-2i+1}; these differ unless i=j. Example: X a 4-fold, i=2, v=i_*w with w∈H^1(D) for a divisor D; then *v is supported on D, hence lies in N^1H^5, not in N^2H^5. Thus (14) is unsupported and the non-degeneracy argument in Corollary 7.8 collapses.","section":"§7, Eq. (14)"},{"comment":"The stated equivalences Λ(X,i)+Λ(X,j) ⇔ GC(X,i)+GC(X,j) and Λ(X,i)⇔GC(X,i) when d+1=2i are not proved in §7. The text proves only the implication Λ⇒GC, and that proof depends on the unproved equation (14). Remark 7.10 in fact says that Condition (Λ,X,i) gives only one inclusion J^j_alg⊂J^i_alg^*, and that the reverse inclusion requires Condition (Λ,X,j). Either supply the missing converse argument or weaken the summary to state only the implication proved.","section":"Summary 7.11(4)"}],"minor_comments":[{"comment":"The title contains spacing/typo artifacts ('INCIDENCE EQUIV ALENCE, A SUR VEY.') that should be cleaned; same for the abstract.","section":"Title and abstract"},{"comment":"Typo 'whch' should be 'which'.","section":"§5.2"},{"comment":"The expression 'G(X,2)' should be 'GC(X,2)'.","section":"Summary 7.11(3)"},{"comment":"The verification of Λ(X,2) is very terse: the assertion 'H^1_prim,alg=H^1(X)_Q by Abel's theorem' is not a standard formulation of Abel's theorem and needs a precise reference or proof. Once the definition of H_alg is corrected, this point becomes non-obvious and should be expanded.","section":"Remark 7.6"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a survey with an original proof (Corollary 7.8) that is currently not supported because of the coniveau-indexing error and the unproved star-preservation claim. I do not see a counterexample to the intended statement, and the defects are likely repairable by reindexing H_alg and giving a correct proof of (14) (or by replacing the star argument with a Poincaré-duality argument). I therefore recommend major revision rather than rejection. The equivalence claims in Summary 7.11(4) should be scaled back unless a proof is supplied."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the survey part is worth having, but Corollary 7.8 — the paper's new result — is not proven. The stress-test note is on target. As written, H_alg^{2i-1} := N^i H^{2i-1} is zero under the paper's own coniveau convention: a subspace of codimension i has real codimension at least 2i, which exceeds 2i-1. So the definition must be reindexed to N^{i-1} or the maximal level-≤1 Hodge substructure. This is load-bearing, not cosmetic, because Condition (Λ,X,i) and Lemma 6.3 both treat this space as nontrivial.\n\nEven after reindexing, equation (14) does not follow from the one-line appeal to Weil's p.76 formula. The star operator, expressed via the Weil operator C and powers of L, is not shown to preserve the rational algebraic subspace. C does not preserve rational classes in general, and L does not raise coniveau. Concretely, on a 4-fold with i=2, the Gysin image of a divisor class lies in the corrected H^3_alg, but its star, supported on the same divisor, lies in N^1 H^5, not N^2 H^5. So *v need not land in the claimed target. The non-degeneracy argument for the cup product, and with it Corollary 7.8, collapses.\n\nWhat the paper does well: it gives a readable survey of incidence equivalence, the standard conjectures, Abel-Jacobi maps, and the recent height-pairing motivation. It is honest about the gap in Murre's 1985 proof, and the idea of using Hodge-star positivity to prove nondegeneracy is a reasonable line to try. But the new proof is only a sketch, and the key step is unsupported as written.\n\nFor a reader, the survey chapters are a fine entry point. The new theorem should not be cited as established. I would send this to peer review because the topic matters and the criterion is plausible; a referee could push the author to either prove (14) properly or downgrade the claim to a conditional statement. The current version needs substantial revision.","headline":"Useful survey, but the advertised new criterion is built on a misindexed definition and an unproved star-operator step.","tokens_in":17814,"tokens_out":9196,"would_cite":false,"duration_ms":90290,"reading_group":"maybe","serious_thinker":"no","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14C25","14C30","14K30"],"pacs":[],"model":"deepseek-v4-flash","headline":"This survey argues that a single Hodge-theoretic condition — all primitive Lefschetz components of algebraic odd cohomology being algebraic — makes incidence equivalence coincide with Abel-Jacobi equivalence for codimension-i cycles algebra","keywords":["incidence equivalence","Abel-Jacobi equivalence","intermediate jacobians","Lefschetz decomposition","generalized Hodge conjecture","algebraic cycles","Hodge star operator","height pairing"],"falsifier":"For a smooth projective variety X and an index i satisfying condition (Λ, X, i), compute the cup-product pairing H^{2i-1}_alg(X) × H^{2d-2i+1}_alg(X) and check non-degeneracy; finding a nonzero class v in H^{2i-1}_alg(X) whose Hodge star is not in H^{2d-2i+1}_alg(X), or whose self-intersection v·(∗v) is zero, would refute equation (14) and invalidate the proof of the corollary.","tokens_in":16737,"feed_emoji":"🔗","tokens_out":9659,"duration_ms":78834,"temperature":0.7,"pith_summary":"This survey revisits a question from the early 1970s: for cycles algebraically equivalent to zero, is the geometrically defined incidence equivalence the same as the transcendentally defined Abel-Jacobi equivalence? Its main new claim is that a positive answer in codimension i follows from a Hodge-theoretic condition: every primitive component in the Lefschetz decomposition of every algebraic class in H^{2i-1} must itself be algebraic. The paper proves this implication in Corollary 7.8 using the Hodge star operator and a positivity statement, and it shows the condition holds in important cases: codimensions 1, 2, and d, abelian varieties, and odd-dimensional complete intersections whose odd cohomology is concentrated in the middle degree. If the two equivalences coincide, Abel's theorem for divisors gains a higher-codimension analogue, and the asymptotic invariant of the archimedean height pairing becomes a purely geometric quantity. The survey also gathers known results and open connections to the generalized Hodge conjecture.","feed_headline":"Algebraic primitive classes align incidence and Abel-Jacobi equivalence","feed_subtitle":"A geometric zero-cycle condition may match the transcendental one, extending Abel's theorem to higher codimensions.","key_machinery":"The argument runs through the Hodge star operator and the Lefschetz decomposition. A classical formula from the theory of Kähler varieties expresses the Hodge star of a primitive class via the Weil operator and the Lefschetz operator; under condition (Λ, X, i) this formula implies that the star of an algebraic class in H^{2i-1} belongs to the algebraic part of H^{2d-2i+1} and has positive self-intersection unless the class is zero. That makes the cup-product pairing between H^{2i-1}_alg and H^{2d-2i+1}_alg non-degenerate on the first factor, and a known criterion — based on biextensions — turns this non-degeneracy into the equality of incidence and Abel-Jacobi equivalence.","core_discovery":"The central new result, Corollary 7.8, states: if every primitive component in the Lefschetz decomposition of every class in H^{2i-1}_alg(X, Q) is itself algebraic — the condition the author names (Λ, X, i) — then for codimension-i cycles algebraically equivalent to zero, incidence equivalence coincides with Abel-Jacobi equivalence after tensoring with Q. The condition is shown to hold for codimensions 1, 2, and d, for abelian varieties because the Λ-operator is algebraic there, and for odd-dimensional complete intersections whose odd cohomology is concentrated in the middle degree. In the codimension-2 case the paper repairs a gap in an earlier published proof.","pith_inferences":["The condition (Λ, X, i) is formulated as an assumption on primitive components; testing it on examples beyond complete intersections and abelian varieties — for instance, low-degree hypersurfaces with richer odd cohomology — would map the true boundary of the theorem.","The positivity assertion underlying (14) is reminiscent of the Hodge-Riemann relations; if it could be derived from those relations directly, the dependence on the explicit star-operator formula might be removed, and the method might extend to variations of Hodge structure.","A single algebraic class whose Hodge star is not algebraic would pinpoint exactly where the geometric-transcendental bridge breaks; hunting for such a class may be more tractable than attacking the generalized Hodge conjecture head-on.","The same correspondence-based strategy could be applied to other cycle-equivalence relations defined by kernels of correspondences, suggesting a general principle: Hodge-theoretic non-degeneracy equals geometric equivalence."],"forward_implications":["A uniform proof of the codimension-2 case replaces a previously incomplete argument, and the same Hodge-theoretic criterion covers all codimensions.","For abelian varieties, incidence and Abel-Jacobi equivalence coincide for cycles of every codimension.","If the generalized Hodge conjecture holds in the relevant odd degree, the coincidence follows for any smooth projective variety.","On odd-dimensional complete intersections of dimension 2m+1 with no odd cohomology outside the middle, the two equivalences agree for codimension m+1 cycles.","When the coincidence holds, the leading asymptotic coefficient of the archimedean height pairing is the local geometric height pairing, a genuine invariant of the cycles."],"fun_headline_variants":["New condition aligns incidence and Abel-Jacobi equivalence","Abel's theorem extended to higher dimensions via incidence","Survey shows when zero cycles match: incidence equals Abel-Jacobi","Primitive class condition unifies incidence and Abel-Jacobi","Codim-2 gap repaired in incidence Abel-Jacobi link"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The proof of Corollary 7.8 rests on equation (14), which asserts that the Hodge star of an algebraic class in H^{2i-1} is again algebraic and has positive self-intersection; this is stated in one sentence from a classical formula, and if the star fails to preserve algebraicity the non-degeneracy of the cup-product pairing is not established.","fun_headline_variants_meta":{"raw":{"variants":["New condition aligns incidence and Abel-Jacobi equivalence","Abel's theorem extended to higher dimensions via incidence","Survey shows when zero cycles match: incidence equals Abel-Jacobi","Primitive class condition unifies incidence and Abel-Jacobi","Codim-2 gap repaired in incidence Abel-Jacobi link"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000971,"raw_usage":{"total_tokens":3920,"prompt_tokens":654,"completion_tokens":3266,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":398,"completion_tokens_details":{"reasoning_tokens":3186}},"tokens_in":398,"tokens_out":3266,"duration_ms":22446,"temperature":1.0,"reasoning_tokens":3186,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T05:22:45.654726+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a smooth projective variety X and an index i satisfying condition (Λ, X, i), compute the cup-product pairing H^{2i-1}_alg(X) × H^{2d-2i+1}_alg(X) and check non-degeneracy; finding a nonzero class v in H^{2i-1}_alg(X) whose Hodge star is not in H^{2d-2i+1}_alg(X), or whose self-intersection v·(∗v) is zero, would refute equation (14) and invalidate the proof of the corollary.","supporting_citations":[],"review_version":1}