Halting Witness · Pre-Analysis Gate
H0
✓ PASS
Domain declared: Pure Mathematics
H1
✓ PASS
Frame: Complex projective geometry, Hodge decomposition
H2
✓ PASS
Dimensional closure: Cohomology degrees type-check
H3
✓ PASS
Mode: Hypothesis. Decisive test specified per open constraint
H4*
◎ N/A
Physics-specific. Mathematics domain: invariant budget replaces c-Spectrum
All Halting Witness gates passed. Analysis proceeds in Hypothesis Mode.
H0 · Domain Primitive Mapping
| Domain |
Pure Mathematics — Algebraic Geometry & Hodge Theory |
| Frame Primitive |
Cohomological frame: H^{p,q}(X) summands of the Hodge decomposition on a smooth complex projective variety X |
| Invariant Primitive |
Rational cohomology class: [Z] ∈ H^{2k}(X, ℚ) — quantity preserved under algebraic equivalence of cycles |
| Topological Closure |
The Hodge class condition: [Z] ∈ H^{k,k}(X) ∩ H^{2k}(X, ℚ). This is the closure analog — the cycle's cohomology class must be of pure bidegree |
| Causal Arrow |
Geometry → Cohomology: algebraic structure of X determines which cohomology classes are representable by algebraic cycles. The conjecture asks whether the causal arrow is also reversible: cohomological data → algebraic cycle existence |
| Attack Classes Active |
A1 (Frame/type confusion), A2 (Invariant violation), A3 (Circularity/tautology), A4 (Post-hoc prediction), A5 (Non-discrimination) + Domain slot: A6-M (Mathematical domain — algebraic vs analytic category confusion) |
The Conjecture · Clay Millennium Problem
On a non-singular complex projective algebraic variety, any Hodge class is a rational linear combination of classes of algebraic cycles.
Let X be a smooth complex projective variety of complex dimension n. For each integer k, the natural map
CH^k(X) ⊗ ℚ → Hdg^k(X) := H^{k,k}(X) ∩ H^{2k}(X, ℚ)
is surjective. Equivalently: every rational cohomology class of type (k,k) is the class of an algebraic cycle with rational coefficients.
Three Independent Validators · Parallel Execution
Derived
Hodge Decomposition Closure
The Hodge decomposition H^n(X,ℂ) = ⊕_{p+q=n} H^{p,q}(X) is geometrically closed: the Kähler condition on X forces this splitting to be orthogonal and complete. The geometric closure condition is satisfied by assumption (smooth projective variety). PASS π-1
Open
Cycle Class Map Injectivity
The cycle class map cl: CH^k(X) → H^{2k}(X, ℤ) is well-defined and natural. The fundamental class of any algebraic cycle lands in H^{k,k} — this direction is proved (Lefschetz, Chern–Weil). The geometric question: does every H^{k,k} class have a cycle closing back to it? This inverse closure is the conjecture. OPEN π-2
Pass
Mode Transition k=1
For k=1, Lefschetz (1,1) theorem confirms the conjecture: every integral H^{1,1} class is the class of a divisor. Geometric closure holds at this mode. The conjecture fails to be derived for k ≥ 2. Mode boundary declared. PASS k=1
Derived
Invariant Budget: Rational Coefficients
The conjecture uses rational coefficients deliberately. Over ℤ, torsion obstructions arise (Atiyah–Hirzebruch). The rational coefficient requirement is an invariant budget choice: it discards torsion, purchasing tractability at the cost of integral precision. This restriction is not ad hoc — it matches the invariant preserved by algebraic equivalence. PASS E-1
Open
Invariant Source: Transcendental Leakage
The space Hdg^k(X) may contain classes with no algebraic source. Transcendental cycles (those not defined over ℚ̄) represent genuine "energy" unaccounted in the algebraic budget. Whether such classes exist on projective varieties is the core invariant-source question. No example is known; none ruled out for k ≥ 2. OPEN E-2
Derived
Hard Lefschetz Constraint
Hard Lefschetz forces the Hodge numbers to satisfy h^{p,q} = h^{q,p} and constraints on h^{k,k}. This constrains the invariant budget from above. Any valid cycle class must fit within the dimensions forced by Poincaré duality. No overcount is possible — the upper bound is derived. PASS E-3
Pass
T-1 · Dependency Acyclicity
Dependency graph: Kähler geometry → Hodge decomposition → Hodge class definition → conjecture statement → alleged algebraic cycle. No cycle. The conjecture does not use its own conclusion. Premises are independently established. T-1 PASS
Pass
T-7 · Description vs Explanation
The Hodge condition [Z] ∈ H^{k,k} ∩ H^{2k}(X,ℚ) is a description of a cohomological property. The conjecture adds a causal arrow: this property is caused by algebraic geometry, i.e., by the existence of an actual cycle. T-7 requires the conjecture to add at least one causal arrow beyond observation — it does. T-7 PASS
Open
A6-M · Category Confusion Risk
Primary risk: conflating analytic cycles (defined by complex analysis) with algebraic cycles (defined by polynomial equations). The Chow group CH^k(X) lives in the algebraic category; the Hodge class lives in the analytic category. The conjecture bridges them — this bridge is the content. Attack A6-M: any proposed proof must not silently cross categories. MONITOR A6-M
Hypothesis Mode · Causal Hypotheses Declared Before Testing
Causal Claim
Every Hodge class arises from the arithmetic of the variety's defining equations via the regulator map K_{2k}(X) → H^{k,k}(X). The algebraic K-theory of X provides the cycle factory; the Chow group is the shadow of K-theory at the level of cohomology.
Competing Hypothesis
H-B (Motivic): Hodge classes arise from the motivic cohomology H^{2k}_{\mathcal{M}}(X, ℚ(k)). H-A and H-B are related but differ in what structure is asserted as primitive. H-B is stronger; H-A is a shadow of H-B.
Prediction (committed before test)
P-A1: For abelian varieties, every Hodge class is algebraic (consistent with Tate conjecture over finite fields). P-A2: Counterexamples, if they exist, must arise on varieties where K-theory and Chow groups diverge — expected in degree ≥ 4.
Status After Known Tests
Confirmed for k=1 (Lefschetz). Confirmed for abelian varieties (Piatetski-Shapiro, Deligne 1982 — Hodge classes on abelian varieties are algebraic). Not confirmed for general k ≥ 2. No counterexample known.
Causal Claim
Grothendieck's Standard Conjectures — particularly the Lefschetz standard conjecture — would imply the Hodge conjecture as a consequence. The causal chain: algebraic correspondences generate all cohomological operations, leaving no room for non-algebraic Hodge classes.
Structural Status
The Standard Conjectures are themselves unproved. H-B therefore defers to a harder open problem. Using H-B to prove the Hodge conjecture would require first proving the Standard Conjectures — this is an upward, not downward, reduction. T-2 flag: H-B inverts the natural complexity ordering.
Prediction
P-B1: Over finite fields (where Tate conjecture = Hodge analog), the arithmetic version is proved (Tate 1966). This gives conditional confidence in the complex case. P-B2: Any proof of the Standard Conjectures would immediately resolve the Hodge conjecture.
Red Kernel Note
A4 risk: arguments that look like proofs of H-B often post-hoc adapt to known cases. The decisive test must distinguish H-B from H-C (Transcendental) in degree ≥ 2 before the data is examined. H-B status: OPEN — no resolution achieved.
Causal Claim
The conjecture is false. There exist smooth projective varieties bearing Hodge classes that are transcendental — not in the image of the cycle class map. The causal mechanism: Hodge symmetry imposes cohomological constraints but does not force algebraic representability.
Evidence Against H-C
No counterexample has been found in 70 years of search. The Lefschetz (1,1) theorem, Deligne's abelian variety result, and all verified cases support H-A/H-B. The absence of counterexamples is weak evidence only — it constrains where a counterexample can hide, not whether it exists.
Decisive Discriminator Against H-A
H-C predicts that a careful construction of a variety of type F₄ (certain Shimura varieties) might yield a Hodge class with no algebraic source. The Moonen–Zarhin theorem constrains but does not eliminate this possibility. This is the current research frontier.
ID
Prediction Content
From
Status
P-1
Hodge conjecture holds for divisors (k=1) on any smooth projective variety
Derived
CONFIRMED
P-2
All Hodge classes on abelian varieties are algebraic
H-A
CONFIRMED
P-3
Integral Hodge conjecture is false: torsion Hodge classes exist with no integral cycle
E-1
CONFIRMED
P-4
Tate conjecture over finite fields (arithmetic analog) holds for divisors
H-B
CONFIRMED
P-5
For k ≥ 2 on a general variety, the surjectivity of CH^k ⊗ ℚ → Hdg^k is unresolved
All
OPEN
P-6
A counterexample, if it exists, must live in dimension ≥ 4 and involve a non-abelian variety
H-C
UNTESTED
Open Constraints · Structural Defect Map
The core structural gap: the Hodge class condition is an
analytic condition (the class decomposes as bidegree
(k,k) under the Dolbeault decomposition), but the cycle class map maps from
algebraic geometry. No general mechanism is known that forces analytic symmetry to imply algebraic representability. This is the precise structural defect.
Decisive test: Construct a smooth projective 4-fold X over ℂ with h^{2,2}(X) > rank(CH^2(X)⊗ℚ) and identify a specific class in the gap. If such a class is Hodge but not algebraic, the conjecture fails. If every such class is forced to be algebraic by an additional argument — the mechanism must be exhibited.
The Lefschetz (1,1) theorem exploits exponential sheaf sequence:
0 → ℤ → 𝒪_X → 𝒪*_X → 0 gives a long exact sequence connecting divisor classes to Hodge classes. For
k ≥ 2, the analogous tool — the sheaf of holomorphic
k-forms — does not yield an integer structure. The obstruction is intrinsic: sheaf cohomology in degree
k loses the link to integral data that powered
k=1. The structural constraint is that no analog of the exponential sequence exists for higher degrees.
Decisive test: Identify whether a derived category replacement — the Hodge-to-de Rham spectral sequence refined to integral data — can reconstruct the exponential sequence structure at k ≥ 2. Failure to find such a replacement by exhaustive inspection of known spectral sequences confirms the structural obstruction is genuine.
A Hodge class varies flatly in a family of varieties (Deligne's principle of absolute Hodge classes). This
rigidity might imply algebraic origin — since algebraic classes also vary flatly. Deligne proved all known Hodge classes are
absolutely Hodge, and conjectured that all absolutely Hodge classes are algebraic. This intermediate conjecture (AHC) is weaker than HC and may be more tractable.
Decisive test: Prove or disprove that absolutely Hodge classes on Shimura varieties of exceptional type (G₂, F₄, E₆, E₇, E₈) are algebraic. Goursat–Kolár–Schoen constructions target this. A single verified non-algebraic absolutely Hodge class would refute HC entirely.
The Chow group
CH^k(X) is known to be
infinite-dimensional over
ℂ in general (Mumford 1969 for surfaces). The Hodge group
Hdg^k(X) is finite-dimensional. Surjectivity does not require the Chow group to be simple — only that its image under the cycle class map covers
Hdg^k. The complexity mismatch is not a direct obstruction but makes the problem computationally inaccessible via brute force. The surjectivity question must be answered structurally, not by enumeration.
Deferral condition: construct a first example of a Hodge class for which a constructive algebraic cycle can be exhibited by computer algebra — in degree k=2 on a specific 3-fold. This would not prove HC but would demonstrate that the cycle class map is surjective in a case where its surjectivity was previously unknown.
A1
Frame violation — algebraic vs analytic category confusion. The conjecture claims a class in the analytic world (Hodge decomposition, defined via differential forms) equals a class from the algebraic world (Chow group, defined by polynomial equations). Any purported proof must explicitly construct the bridge. Merely showing both objects exist in H^{2k}(X,ℚ) without demonstrating they are the same element is a frame boundary crossing.
Verdict: Real attack. All proposed proofs to date have failed here. Structural defect A1 is active.
A3
Circularity attack — using Hodge structure to define algebraicity. Some approaches attempt to define algebraic cycles as those whose cohomology class is Hodge, then prove all Hodge classes are algebraic. This is circular: algebraicity is the premise used in the definition. Attack A3 targets this — the dependency graph has a cycle at the node "algebraic cycle."
Verdict: Real risk in motivic approaches. T-1 acyclicity check flags this pattern. No known proof attempt has fully avoided it.
A4
Post-hoc prediction — the abelian variety case. Deligne's theorem on Hodge classes for abelian varieties (1982) is sometimes cited as strong evidence for HC. Attack A4: this theorem uses CM (complex multiplication) structure specific to abelian varieties. Claiming it evidences the general case requires showing the mechanism generalizes — which has not been done. The prediction "all abelian variety Hodge classes are algebraic" was not cleanly committed before Deligne's proof.
Verdict: Partial flag. Deligne's theorem is independently valuable but does not discriminate H-A from H-B in the general case.
A5
Non-discrimination — HC and ¬HC produce identical predictions in all tested cases. Both the conjecture and its negation are consistent with all known computational evidence (no counterexample found; no proof). The hypotheses currently have zero discrimination at the level of computable examples in degree k ≥ 2.
Verdict: Active failure — A5 is the primary reason the conjecture remains open. Designing a discriminating experiment is OC-1's decisive test.
A6-M
Domain-specific — GAGA principle abuse. Serre's GAGA theorem equates coherent algebraic and analytic sheaves on projective varieties. Attempts to use GAGA to convert analytic Hodge classes into algebraic cycles are tempting but invalid: GAGA operates on sheaves of functions, not on homology classes. The cycle class map is not a sheaf morphism.
Verdict: Known false path. Multiple early attempts failed here. Attack A6-M correctly identifies GAGA as a frame boundary that cannot be crossed without additional structure.
25–54
◄ STRUCTURAL ISSUES · Current Position
Penalty breakdown: A1 frame violation (active) +25 · A5 non-discrimination (active) +10 · A3 circularity risk +10 · A4 partial flag +7 · A6-M domain risk (known, manageable) +0 (flagged only).
The score reflects the current state of knowledge — the conjecture has structural defects in its evidential base, not in its formulation.
T-6 · Compression Assessment
Hdg²(X) := H^{k,k}(X) ∩ H^{2k}(X, ℚ)
The Hodge class definition compresses: (1) complex differential geometry (Dolbeault decomposition), (2) algebraic topology (singular cohomology over ℚ), and (3) Kähler geometry — into a single containment condition. Compression ratio is high. The conjecture claims this compressed condition is algebraically complete — that no additional analytic data is needed beyond what the Chow group already supplies. The explanatory target is precise; the mechanism is missing.
Decisive Test · Most Important Open Constraint Resolution
Target Constraint
OC-1: The Algebraic-Analytic Bridge — no general mechanism exhibited
Construction
Take a smooth projective 4-fold X (e.g., a complete intersection of type (2,2,2) in ℙ⁶) with H^{2,2}(X) ≠ 0 and b_4(X) large. Compute Hdg^2(X) via period integrals. Attempt to exhibit an explicit algebraic cycle for each generator.
Confirmation Condition
If every generator of Hdg^2(X) is matched by an explicit element of CH^2(X) ⊗ ℚ with matching cycle class: H-A/H-B supported, OC-1 partially closed, decisive positive evidence accumulated.
Refutation Condition
If a Hodge class is found for which no algebraic cycle can be exhibited despite exhaustive search in the Chow group (via computer algebra), and if the failure is structural (not just computational): H-C becomes viable. Conjecture is false candidate confirmed.
What distinguishes this
This test discriminates H-A from H-C without circular reasoning. It commits to an algebraic construction before examining the Hodge data for that specific variety. It satisfies EHP logging requirements: observation → hypothesis → prediction → test → outcome logged.
Σ-Kernel Session Record · HC-001 · Appendix
A.1 Domain Declaration (H0)
Domain: Pure Mathematics · Primitive: Algebraic geometry over ℂ · Causal arrow: algebraic structure → cohomological data (direct); HC claims the reverse arrow also holds · Attack classes: A1, A2, A3, A4, A5, A6-M active
A.2 Hypothesis Status at Session End
H-A
IN-PROGRESS — Confirmed for k=1 and abelian varieties. Not confirmed for general k ≥ 2. No mechanism identified.
H-B
OPEN — Standard Conjectures unproved. H-B defers to a harder problem (T-2 inversion flagged). Retains logical coherence.
H-C
OPEN — No counterexample found in 70 years. Not refuted. Remains a logical possibility. Constrained to dim ≥ 4, non-abelian varieties.
A.3 ORE Log (Ontological Revision Events)
ORE-1
Integral Hodge conjecture initially considered equivalent to rational HC. Revised: Atiyah–Hirzebruch counterexamples establish strict separation. Prior validator outputs on integral HC invalidated. Rational HC treated as independent target.
ORE-2
Early framing treated HC as a statement about the image of a single map. Revised: HC is a surjectivity claim requiring the map to have no cokernel. The correct framing requires stating what the target space Hdg^k(X) is — not just that a map exists. H0 domain declaration updated to include this distinction.
A.4 T-1 Circularity Audit — Dependency Trace
Smooth projective variety X [given] → Kähler metric [exists by Kodaira] → Hodge decomposition [derived: Hodge theorem] → Hodge class definition [defined] → Cycle class map cl: CH^k → H^{2k} [defined, direction proved] → Surjectivity claim [asserted by HC] → Algebraic cycle existence [open] → No ζ-function, no circular reasoning. T-1: PASSED.