Progress · independent proof diagram
Products of curves and surfaces
Subject to the product-model theorem, the construction covers every complementary pair of homologically trivial Chow groups on a product of arbitrary curves and surfaces, and on the stated Chow-motive summands of such products.
Conditional theorem with an additional model hypothesis · Source edition v0.5.1
Domain of the result
, where , the are smooth projective curves, the are smooth projective surfaces, and has potentially good projective reduction at every finite place. One may take to be a point.
Read the diagrams from top to bottom. Select a numbered connection for its exact hypotheses and deduction, or follow a node to its source. Analytic, model and comparison assumptions retain their stated scope.
How to read the logic
Read from top to bottom. Parallel premises meet at a numbered connection, then lead to a conclusion. Select a connection to read its hypotheses and deduction; select a node to read the linked statement.
The full scope of each occurrence is recorded below its diagram. A proved conditional implication and an open geometric premise keep their separate statuses.
LOGIC 01
Products of curves and surfaces
Low-dimensional endpoints and potential good reduction cover the boundary cases. For nonempty products of dimension at least three, the model inverse, odd weights and native descent give the conditional local argument.
Scope. All Chow groups have rational coefficients. Vʳ means geometric homological triviality, component by component. Local assertions use the original finite valued field K/ℚₚ; global assertions use a number field F, n=dim X and 1≤r≤n, s=n+1−r.
Scroll inside the frame to follow a large branch. “Fit width” shows the whole layout; “Readable size” restores the text to 100%.
Read the 6 connections
Each numbered card gives the exact inputs, the deduction and its sources.
01AND · Use AᴴGᴴAᴴ=Aᴴ on FULL cohomology
Premises and stated hypotheses
- HypothesisThe product-model inverse is an additional hypothesis
G is a rational Chow-correspondence matrix of codimension d−1 and A has codimension d+1 on d-dimensional components. Its identity must hold in every cohomological degree and Tate twist, including transcendental and exceptional summands.
Scope & sources
How the deduction works
If p(x)∈ker Aᴺ, x♭=x−GᴴAᴴx has the same numerical image and Aᴴx♭=0. If α is algebraic and lies in im Aᴴ, then Gᴴα is algebraic and AᴴGᴴα=α.
Conclusion / target
A=σρ and the actual algebraic G preserve numerical equivalence and actual algebraic class spans.
Scope & sourcesEvidence and further reading
Back to connection 01 in the diagram ↑02AND · The dimension range + the divisor and zero-cycle constructions
Premises and stated hypotheses
- Hypothesis1≤n≤2 and 1≤r≤n
For a curve r=1=n; for a surface the only positive degrees are r=1 and r=2=n.
Scope & sources - ConditionalOb¹X vanishes on homologically trivial divisors
The linked complete proof passes from geometric homological triviality to rational algebraic triviality and then uses the independent compact/Poincaré construction under (A1), (A2) and (R). It produces a global primitive for the prescribed curvature, as well as exact-support strong G.
Scope & sources - ConditionalObⁿX vanishes on homologically trivial zero-cycles
The linked complete proof uses degree zero on each geometric component, geometric point-difference families, and the same compact/Poincaré theorem under (A1), (A2) and (R). The conclusion concerns the original actual model curvature, not arbitrary atomic forms.
Scope & sources
How the deduction works
When 1≤n≤2 every 1≤r≤n is either r=1 or r=n. The two endpoint theorems therefore annihilate the original obstruction in every required degree; no AK19 model criterion is used.
Conclusion / target
This branch applies to any of the permitted products with 1≤n≤2, including m=0. Both endpoint proofs retain their actual-form hypotheses.
Scope & sourcesEvidence and further reading
- r20dp:green, first paragraph of the proof
- The endpoint statements
- r12p:picard-products; r12p:top-degree; r12p:geometric-algebraic
03AND · Empty-product scope + a potential smooth projective good model + native vanishing
Premises and stated hypotheses
- Hypothesism=0, X=T and n≥1
There are no curve or surface factors. The degree range remains 1≤r≤n; the zero-dimensional coefficient contributes no such degree.
Scope & sources - HypothesisT has potential smooth projective good reduction
Choose a finite valued L/K and a smooth projective model of T_L. For the global height conclusion this hypothesis is required at every finite place, separately.
Scope & sources - ConditionalThe finite good-cover proof gives ObʳX([Z])=0
The complete branch below compares the original pulled curvature with a zero good-model curvature, then descends the original permitted line sums. It uses the stated analytic interfaces and does not trace an arbitrary primitive.
Scope & sources
How the deduction works
Take the finite projection X_L→X after a valued extension admitting a smooth projective model. Apply the structured-primitive and composite-trace proof with full degree [L:K] and the original absolute value.
Conclusion / target
For m=0 the product is X=T. The finite good-cover argument below proves vanishing of the original model-curvature class over K, so it supplies the needed domain statement directly.
Scope & sourcesEvidence and further reading
- Empty products are explicitly permitted
- r11o:finite-good-vanishing
- Potential good reduction and native finite degree
04AND · m≥1 and n≥3 + fixed witness + exactness + odd weights + comparison
Premises and stated hypotheses
- Hypothesism≥1 and n≥3
DP20 states the product theorem for m≥1 and handles n≤2 first by endpoints. Its proof invokes the AK19 model criterion only after assuming n≥3.
Scope & sources - HypothesisOne common semistable product witness
After finite extension take projective strictly semistable alterations of the curve/surface factors and a smooth projective model of T. Repair later ramified base changes before taking the ordered staircase resolution. Keep all connected intersections.
Scope & sources - ConditionalNumerical-kernel lifting and actual-span saturation
A=σρ and the actual algebraic G preserve numerical equivalence and actual algebraic class spans.
Scope & sources - HypothesisOdd-weight vanishing on the same witness
Weight–monodromy for all cohomological degrees of curves and surfaces, the pure unramified good coefficient, and Künneth give the companion odd-weight condition. The comparison from this condition to the actual inertia term is retained.
Scope & sources - HypothesisThe supplied model comparison and native descent
Use the stated actual-obstruction, numerical-model, dominant-detection and finite-trace interfaces; their analytic audit is unfinished. Retain full generic and field degrees and the original absolute value. The final form uses the original specified support and all actual tests.
Scope & sources
How the deduction works
Only this range invokes AK19. The algebraic calculation is on the actual ordered semistable witness; dominant detection and the full finite field degree descend its obstruction before the original curved pair is corrected.
Conclusion / target
This is the DP20 branch for a nonempty curve/surface product of dimension at least three. All original model and analytic assumptions remain in force.
Scope & sourcesEvidence and further reading
Back to connection 04 in the diagram ↑05OR · Low dimension OR empty product OR the higher-dimensional model route
Alternative complete routes
- ConditionalAll positive degrees are endpoint degrees when n≤2
This branch applies to any of the permitted products with 1≤n≤2, including m=0. Both endpoint proofs retain their actual-form hypotheses.
Scope & sources - ConditionalThe coefficient alone has zero obstruction
For m=0 the product is X=T. The finite good-cover argument below proves vanishing of the original model-curvature class over K, so it supplies the needed domain statement directly.
Scope & sources - ConditionalThe fixed product model gives native obstruction zero
This is the DP20 branch for a nonempty curve/surface product of dimension at least three. All original model and analytic assumptions remain in force.
Scope & sources
How the deduction works
The branches together cover the permitted product domain. The AK19 route is used only when m≥1 and n≥3. Empty products are handled by native finite good-cover obstruction vanishing, not by a converse from existence of a Green form.
Conclusion / target
Write m for the number of curve/surface factors and n=dim X. For 1≤r≤n use a complete branch in its stated range. Run that branch anew for every newly specified cycle after every finite K′/K; no descent of the new class to K is assumed. If n=0 there is no codimension in the stated range.
Scope & sourcesEvidence and further reading
- Product scope: empty products and new classes are permitted
- r20dp:green: the dimension split and the m≥1 hypothesis
- r11o:finite-good-vanishing: obstruction zero for the empty product
06AND · The all-degree local result + the global height hypotheses
Premises and stated hypotheses
- ConditionalAll homological product classes have zero obstruction over K
Write m for the number of curve/surface factors and n=dim X. For 1≤r≤n use a complete branch in its stated range. Run that branch anew for every newly specified cycle after every finite K′/K; no descent of the new class to K is assumed. If n=0 there is no codimension in the stated range.
Scope & sources - HypothesisAdmitted-domain reciprocity at every finite completion
The global input supplies independence, reciprocity, archimedean terms and rational-equivalence descent. Potentially good projective reduction of T is required separately at each finite place and persists after finite extension.
Scope & sources
How the deduction works
Apply the local argument in r and s=n+1−r at every finite completion; both whole homologically trivial groups are then admitted.
Conclusion / target
The curves and surfaces are arbitrary smooth projective geometrically integral varieties. T may be a point; otherwise it has potentially good projective reduction at every finite place.
Scope & sourcesEvidence and further reading
Back to connection 06 in the diagram ↑All 18 nodes: scope, status and sources
ConditionalAll complementary pairings on T × ∏Cₐ × ∏Sᵦ
The curves and surfaces are arbitrary smooth projective geometrically integral varieties. T may be a point; otherwise it has potentially good projective reduction at every finite place.
Back to this node in the diagram ↑ConditionalAll homological product classes have zero obstruction over K
Write m for the number of curve/surface factors and n=dim X. For 1≤r≤n use a complete branch in its stated range. Run that branch anew for every newly specified cycle after every finite K′/K; no descent of the new class to K is assumed. If n=0 there is no codimension in the stated range.
Back to this node in the diagram ↑ConditionalAll positive degrees are endpoint degrees when n≤2
This branch applies to any of the permitted products with 1≤n≤2, including m=0. Both endpoint proofs retain their actual-form hypotheses.
Back to this node in the diagram ↑Hypothesis1≤n≤2 and 1≤r≤n
For a curve r=1=n; for a surface the only positive degrees are r=1 and r=2=n.
Back to this node in the diagram ↑ConditionalOb¹X vanishes on homologically trivial divisors
The linked complete proof passes from geometric homological triviality to rational algebraic triviality and then uses the independent compact/Poincaré construction under (A1), (A2) and (R). It produces a global primitive for the prescribed curvature, as well as exact-support strong G.
Back to this node in the diagram ↑ConditionalObⁿX vanishes on homologically trivial zero-cycles
The linked complete proof uses degree zero on each geometric component, geometric point-difference families, and the same compact/Poincaré theorem under (A1), (A2) and (R). The conclusion concerns the original actual model curvature, not arbitrary atomic forms.
Back to this node in the diagram ↑ConditionalThe coefficient alone has zero obstruction
For m=0 the product is X=T. The finite good-cover argument below proves vanishing of the original model-curvature class over K, so it supplies the needed domain statement directly.
Back to this node in the diagram ↑Hypothesism=0, X=T and n≥1
There are no curve or surface factors. The degree range remains 1≤r≤n; the zero-dimensional coefficient contributes no such degree.
Back to this node in the diagram ↑HypothesisT has potential smooth projective good reduction
Choose a finite valued L/K and a smooth projective model of T_L. For the global height conclusion this hypothesis is required at every finite place, separately.
Back to this node in the diagram ↑ConditionalThe finite good-cover proof gives ObʳX([Z])=0
The complete branch below compares the original pulled curvature with a zero good-model curvature, then descends the original permitted line sums. It uses the stated analytic interfaces and does not trace an arbitrary primitive.
Back to this node in the diagram ↑ConditionalThe fixed product model gives native obstruction zero
This is the DP20 branch for a nonempty curve/surface product of dimension at least three. All original model and analytic assumptions remain in force.
Back to this node in the diagram ↑Hypothesism≥1 and n≥3
DP20 states the product theorem for m≥1 and handles n≤2 first by endpoints. Its proof invokes the AK19 model criterion only after assuming n≥3.
Back to this node in the diagram ↑HypothesisOne common semistable product witness
After finite extension take projective strictly semistable alterations of the curve/surface factors and a smooth projective model of T. Repair later ramified base changes before taking the ordered staircase resolution. Keep all connected intersections.
Back to this node in the diagram ↑ConditionalNumerical-kernel lifting and actual-span saturation
A=σρ and the actual algebraic G preserve numerical equivalence and actual algebraic class spans.
Back to this node in the diagram ↑HypothesisThe product-model inverse is an additional hypothesis
G is a rational Chow-correspondence matrix of codimension d−1 and A has codimension d+1 on d-dimensional components. Its identity must hold in every cohomological degree and Tate twist, including transcendental and exceptional summands.
Back to this node in the diagram ↑HypothesisOdd-weight vanishing on the same witness
Weight–monodromy for all cohomological degrees of curves and surfaces, the pure unramified good coefficient, and Künneth give the companion odd-weight condition. The comparison from this condition to the actual inertia term is retained.
Back to this node in the diagram ↑HypothesisThe supplied model comparison and native descent
Use the stated actual-obstruction, numerical-model, dominant-detection and finite-trace interfaces; their analytic audit is unfinished. Retain full generic and field degrees and the original absolute value. The final form uses the original specified support and all actual tests.
Back to this node in the diagram ↑HypothesisAdmitted-domain reciprocity at every finite completion
The global input supplies independence, reciprocity, archimedean terms and rational-equivalence descent. Potentially good projective reduction of T is required separately at each finite place and persists after finite extension.
Back to this node in the diagram ↑LOGIC 02
The finite good-cover proof for the coefficient alone
Resolve the prescribed pulled cycle on the good model, obtain zero native curvature, compare the original curvature globally, and descend only the original permitted line sums.
Scope. K/ℚp is finite with its original absolute value. Let X/K be smooth projective geometrically integral, n=dim X≥1, 1≤r≤n, and Z a prescribed rational cycle with zero geometric ℓ-adic class for one ℓ≠p. Assume X_L has a smooth projective model over L° for a finite valued L/K. The actual Chern, numerical detector, structured primitive, minimal trace, Stokes and current-injectivity interfaces remain explicit conditional inputs.
Scroll inside the frame to follow a large branch. “Fit width” shows the whole layout; “Readable size” restores the text to 100%.
Read the 4 connections
Each numbered card gives the exact inputs, the deduction and its sources.
01AND · Finite good model + fixed-model resolutions + every native vertical degree zero + detector
Premises and stated hypotheses
- Hypothesish:Y=X_L→X is finite flat of degree a=[L:K]
X/K is smooth projective geometrically integral and L/K is finite valued with a smooth projective model of X_L. Regard Y and that model natively over K. Flat pullback includes the generic scheme lengths and preserves geometric homological triviality.
Scope & sources - ConditionalResolve the actual flat closures on the fixed smooth model
The smooth projective L°-model is regular and has an ample line bundle. Repeated surjections from sums of negative ample twists terminate in a locally free kernel, giving finite global resolutions of every prescribed component closure without modifying the good model. Their degree-r generic character is h*[Z].
Scope & sources - ConditionalEvery integral vertical r-cycle has zero Chern degree
Smooth proper base change identifies the geometric special Chern class with the zero generic class. Over K° the special fiber is the e(L/K)-fold nilpotent thickening of its L° reduction. Integral cycles lie on that reduction; residue-field change multiplies degrees by f(L/K), so each degree stays zero. Retain all native multiplicities.
Scope & sources - HypothesisThe actual numerical model-Chern detector
This is an analytic input, applied on the projective flat native K°-model and tested against every integral vertical r-cycle. Its output is equality as a K-actual form, not merely vanishing of a current or a cohomology class; the independent analytic audit remains unfinished.
Scope & sources
How the deduction works
Smooth proper base change annihilates the geometric special Chern class. Every integral vertical r-cycle lies on the reduction of the native special fiber; its intersection degree is zero. The native numerical model-Chern detector then annihilates the actual curvature.
Conclusion / target
The numerical detector is applied directly to the good model viewed over K°, including its nonreduced native special fiber. Equality on a smaller collection of L-tests is not substituted for the K-actual assertion.
Scope & sourcesEvidence and further reading
- Fixed smooth-model resolutions and numerical vanishing
- Ramification-thickened native model and full test space
02AND · Zero good-model curvature + a generic Chow-zero difference + constructive primitives
Premises and stated hypotheses
- ConditionalThe resolved pulled cycle has zero K-actual curvature
The numerical detector is applied directly to the good model viewed over K°, including its nonreduced native special fiber. Equality on a smaller collection of L-tests is not substituted for the K-actual assertion.
Scope & sources - Conditionalh*ΩX−Ωgood has generic degree-r Chow class zero
Choose ΩX for the original specified cycle and compare on a common model of Y. Strict Chern pullback and invariance of the original model metrics under refinement preserve h*ΩX. The resolutions have the same generic class h*[Z], and Ωgood=0 by the preceding calculation.
Scope & sources - HypothesisThe global Chow-zero primitive retains its original line sums
The Adams degree-r projector reduces a generic Chow-zero character to a generically K₀-zero virtual class. The finite K₀ relation is resolved by generically acyclic complexes, whose exact generic open is all Y. Their construction gives finite-flat bⱼ on all Y and the permitted global Gⱼ used above. The strict actual Chern and minimal-trace interfaces remain assumptions.
Scope & sources
How the deduction works
On a common projective model compare h*ΩX with the actual character of the good-model resolutions of h*Z. The latter is zero; both characters represent h*[Z], so their difference has generic degree-r Chow class zero. Use the global generically acyclic construction before tracing.
Conclusion / target
Each bⱼ is finite flat on all of Y, each Gⱼ is an original global PL-scalar/model-curvature line sum, DGⱼ=bⱼ*Ωⱼ, and ΣⱼqⱼΩⱼ=h*ΩX. These stronger constructive data permit the later composite trace.
Scope & sourcesEvidence and further reading
- r11o:finite-good-vanishing, construction of βY
- Global primitive for a generic degree-r Chow-zero character
03AND · A structured global primitive upstairs + permitted composite traces downstairs
Premises and stated hypotheses
- ConditionalDβY=h*ΩX with βY=Σⱼqⱼdⱼ⁻¹Tr(bⱼ)(Gⱼ)
Each bⱼ is finite flat on all of Y, each Gⱼ is an original global PL-scalar/model-curvature line sum, DGⱼ=bⱼ*Ωⱼ, and ΣⱼqⱼΩⱼ=h*ΩX. These stronger constructive data permit the later composite trace.
Scope & sources - HypothesisComposite minimal trace, native degree, Stokes and current injectivity
Each h∘bⱼ is finite flat on all X with complete degree a dⱼ. Trace the original Gⱼ in the permitted module, with normalization (a dⱼ)⁻¹ and the original absolute value. The calculation uses every compact K-actual complementary test. No general trace on an arbitrary already constructed βY is assumed.
Scope & sources
How the deduction works
For a=[L:K] define βX=a⁻¹Σⱼqⱼdⱼ⁻¹Tr(h∘bⱼ)(Gⱼ). On each compact actual test η, the pairing is a⁻¹Σⱼqⱼ∫YΩⱼh*η=a⁻¹∫Yh*(ΩXη)=∫XΩXη. Stokes and current injectivity give DβX=ΩX.
Conclusion / target
This is the conditional native finite good-cover route used when m=0. Its proof constructs a global actual primitive of the ORIGINAL ΩX; it does not infer obstruction zero from the existence of another Green form.
Scope & sourcesEvidence and further reading
- r11o:finite-good-vanishing, complete constructive proof
- Native good-model calculation and complete degree
All 13 nodes: scope, status and sources
ConditionalObʳX([Z])=0 over the original valued field
This is the conditional native finite good-cover route used when m=0. Its proof constructs a global actual primitive of the ORIGINAL ΩX; it does not infer obstruction zero from the existence of another Green form.
Back to this node in the diagram ↑ConditionalDβY=h*ΩX with βY=Σⱼqⱼdⱼ⁻¹Tr(bⱼ)(Gⱼ)
Each bⱼ is finite flat on all of Y, each Gⱼ is an original global PL-scalar/model-curvature line sum, DGⱼ=bⱼ*Ωⱼ, and ΣⱼqⱼΩⱼ=h*ΩX. These stronger constructive data permit the later composite trace.
Back to this node in the diagram ↑ConditionalThe resolved pulled cycle has zero K-actual curvature
The numerical detector is applied directly to the good model viewed over K°, including its nonreduced native special fiber. Equality on a smaller collection of L-tests is not substituted for the K-actual assertion.
Back to this node in the diagram ↑Hypothesish:Y=X_L→X is finite flat of degree a=[L:K]
X/K is smooth projective geometrically integral and L/K is finite valued with a smooth projective model of X_L. Regard Y and that model natively over K. Flat pullback includes the generic scheme lengths and preserves geometric homological triviality.
Back to this node in the diagram ↑ConditionalResolve the actual flat closures on the fixed smooth model
The smooth projective L°-model is regular and has an ample line bundle. Repeated surjections from sums of negative ample twists terminate in a locally free kernel, giving finite global resolutions of every prescribed component closure without modifying the good model. Their degree-r generic character is h*[Z].
Back to this node in the diagram ↑ConditionalEvery integral vertical r-cycle has zero Chern degree
Smooth proper base change identifies the geometric special Chern class with the zero generic class. Over K° the special fiber is the e(L/K)-fold nilpotent thickening of its L° reduction. Integral cycles lie on that reduction; residue-field change multiplies degrees by f(L/K), so each degree stays zero. Retain all native multiplicities.
Back to this node in the diagram ↑HypothesisThe actual numerical model-Chern detector
This is an analytic input, applied on the projective flat native K°-model and tested against every integral vertical r-cycle. Its output is equality as a K-actual form, not merely vanishing of a current or a cohomology class; the independent analytic audit remains unfinished.
Back to this node in the diagram ↑Conditionalh*ΩX−Ωgood has generic degree-r Chow class zero
Choose ΩX for the original specified cycle and compare on a common model of Y. Strict Chern pullback and invariance of the original model metrics under refinement preserve h*ΩX. The resolutions have the same generic class h*[Z], and Ωgood=0 by the preceding calculation.
Back to this node in the diagram ↑HypothesisThe global Chow-zero primitive retains its original line sums
The Adams degree-r projector reduces a generic Chow-zero character to a generically K₀-zero virtual class. The finite K₀ relation is resolved by generically acyclic complexes, whose exact generic open is all Y. Their construction gives finite-flat bⱼ on all Y and the permitted global Gⱼ used above. The strict actual Chern and minimal-trace interfaces remain assumptions.
- Adams projection and the generic K₀-zero proof
- r11o:finite-good-vanishing: the stronger structured form of the primitive
HypothesisComposite minimal trace, native degree, Stokes and current injectivity
Each h∘bⱼ is finite flat on all X with complete degree a dⱼ. Trace the original Gⱼ in the permitted module, with normalization (a dⱼ)⁻¹ and the original absolute value. The calculation uses every compact K-actual complementary test. No general trace on an arbitrary already constructed βY is assumed.
Back to this node in the diagram ↑ConditionalThe original GZ−βX is strong G on exactly Xᵃⁿ∖|Z|ᵃⁿ
The global primitive does not add poles or change the prescribed cycle. The original curved pair and every compact actual complementary test remain the ones used in the final identity.
Back to this node in the diagram ↑ConditionalDβX=ΩX on all Xᵃⁿ
Use the native primitive constructed above, not a primitive defined only away from the cycle.
Back to this node in the diagram ↑HypothesisD[GZ]+δZ=[ΩX] on every compact actual test
Use the prescribed collected cycle, its exact support and the supplied local integrability. The curved-pair theorem remains an analytic input.
Back to this node in the diagram ↑LOGIC 03
The complete product-model decomposition
The supplied reconstruction retains the full boundary and every exceptional copy before splitting the transcendental sectors.
Scope. All Chow groups have rational coefficients. Vʳ means geometric homological triviality, component by component. Local assertions use the original finite valued field K/ℚₚ; global assertions use a number field F, n=dim X and 1≤r≤n, s=n+1−r.
Scroll inside the frame to follow a large branch. “Fit width” shows the whole layout; “Readable size” restores the text to 100%.
Read the 4 connections
Each numbered card gives the exact inputs, the deduction and its sources.
01AND · The full boundary tower and the actual fan projection
Premises and stated hypotheses
- Draft · audit pendingEach component is a tower of full-boundary stratum blowups
The staircase rays have height one and maximal-chain differences form an integral basis. Binary staircase induction gives a regular projective strictly semistable model; all original and exceptional boundary rays remain actual neighboring components.
Scope & sources - Draft · audit pendingThe selected-factor open is a literal product
On Uⱼ=Xⱼ∖∂Xⱼ the factor has one vertical ray. Deletion is U_I times the remaining model. No same-label double is hidden above a selected-factor boundary.
Scope & sources
How the deduction works
The map is an isomorphism off the selected factor boundaries. Same-label doubles map with degree one to their deletion doubles; changing a selected label lands in its boundary.
Conclusion / target
W₋I is the actual remaining staircase component, including its exceptional summands. Coordinate deletion sends each chain cone into a chain cone, giving compatible proper birational maps.
Scope & sourcesEvidence and further reading
Back to connection 01 in the diagram ↑02AND · Support localization + boundary annihilation + incidence transport
Premises and stated hypotheses
- Draft · audit pendingThe refined surface projector kills boundary-supported classes
Boundary curves and points carry no image of the selected H²tr projector. Künneth with supports applies after the actual deletion maps.
Scope & sources - Draft · audit pendingTransport every actual double and normal self-intersection
Same-label terms use f′∗f′*=1; changing-label restrictions and Gysin terms vanish. The diagonal is the negative sum of all incident double-divisor classes because the reduced total special fiber is principal.
Scope & sources
How the deduction works
The difference between two deletion pull-push operators is supported over a factor boundary. Its H²tr projector kills it; transposition gives both orders. The same calculation applies to exceptional copies.
Conclusion / target
Choose refined surface projectors pⱼ onto H²tr and set P_I=f_I*(p_I⊠1)f_I∗. They retain the original component labels.
Scope & sourcesEvidence and further reading
Back to connection 02 in the diagram ↑03AND · The full Chow generator decomposition + exact relative-Chow deletion
Premises and stated hypotheses
- Draft · audit pendingu:M=⊕ h(Sα)⊗Lᵗα → h(D), uv=1
Blowup and projective-bundle formulas hold after every smooth projective test product. Original closed product strata, their exact Tate exponents, boundary Chern classes and relative refined intersections are retained.
Scope & sources - Draft · audit pendingA generator with t₂(Xⱼ) factors through fⱼ*
The restricted tower over Uⱼ is Uⱼ×W₋ⱼ. Its actual relative Chow injection agrees there with the exterior product. Localization puts their difference over ∂Xⱼ; a full-degree alteration of each support shows that pⱼ annihilates its cohomology action.
Scope & sources - Draft · audit pendingQ=Q u_ab pr_ab v, with actual correspondence maps
The refined surface complement is Artin–Tate/Picard/Albanese; curve and point factors are abelian type. The resulting u_Qv_Q=Q is a homological retraction, without requiring a new Chow-idempotent lift.
Scope & sources
How the deduction works
Write the component motive using actual stratum blowups and projective bundles. Every generator containing a full-surface t₂ lands in im Pⱼ and is killed by Q; all other source pieces are of abelian type.
Conclusion / target
This is a statement about the residual core after retained t₂ factors have been separated; arbitrary surface motives are not declared abelian type.
Scope & sourcesEvidence and further reading
Back to connection 03 in the diagram ↑04AND · Actual deletion maps + commuting projectors + the residual-core calculation
Premises and stated hypotheses
- Draft · audit pendingf_I:Dᵢ → (∏ⱼ∈I Xⱼ,ᵢⱼ) × W₋I
W₋I is the actual remaining staircase component, including its exceptional summands. Coordinate deletion sends each chain cone into a chain cone, giving compatible proper birational maps.
Scope & sources - Draft · audit pendingP_I=∏ⱼ∈I Pⱼ and APⱼ=PⱼA
Choose refined surface projectors pⱼ onto H²tr and set P_I=f_I*(p_I⊠1)f_I∗. They retain the original component labels.
Scope & sources - Draft · audit pendingQ=∏(1−Pⱼ) is a homological summand of an abelian carrier
This is a statement about the residual core after retained t₂ factors have been separated; arbitrary surface motives are not declared abelian type.
Scope & sources
How the deduction works
The projectors and their commutation are identities on full cohomology of actual rational correspondences. A new Chow-idempotent commutation is not asserted.
Conclusion / target
This is the additional product-model argument. The original DP20 dossier calls its statements candidates pending independent review; the diagram retains that status.
Scope & sourcesEvidence and further reading
Back to connection 04 in the diagram ↑All 11 nodes: scope, status and sources
Draft · audit pendingEach incidence sector is a fixed t₂ tensor times an abelian carrier
This is the additional product-model argument. The original DP20 dossier calls its statements candidates pending independent review; the diagram retains that status.
Back to this node in the diagram ↑Draft · audit pendingf_I:Dᵢ → (∏ⱼ∈I Xⱼ,ᵢⱼ) × W₋I
W₋I is the actual remaining staircase component, including its exceptional summands. Coordinate deletion sends each chain cone into a chain cone, giving compatible proper birational maps.
Back to this node in the diagram ↑Draft · audit pendingEach component is a tower of full-boundary stratum blowups
The staircase rays have height one and maximal-chain differences form an integral basis. Binary staircase induction gives a regular projective strictly semistable model; all original and exceptional boundary rays remain actual neighboring components.
Back to this node in the diagram ↑Draft · audit pendingThe selected-factor open is a literal product
On Uⱼ=Xⱼ∖∂Xⱼ the factor has one vertical ray. Deletion is U_I times the remaining model. No same-label double is hidden above a selected-factor boundary.
Back to this node in the diagram ↑Draft · audit pendingP_I=∏ⱼ∈I Pⱼ and APⱼ=PⱼA
Choose refined surface projectors pⱼ onto H²tr and set P_I=f_I*(p_I⊠1)f_I∗. They retain the original component labels.
Back to this node in the diagram ↑Draft · audit pendingThe refined surface projector kills boundary-supported classes
Boundary curves and points carry no image of the selected H²tr projector. Künneth with supports applies after the actual deletion maps.
Back to this node in the diagram ↑Draft · audit pendingTransport every actual double and normal self-intersection
Same-label terms use f′∗f′*=1; changing-label restrictions and Gysin terms vanish. The diagonal is the negative sum of all incident double-divisor classes because the reduced total special fiber is principal.
Back to this node in the diagram ↑Draft · audit pendingQ=∏(1−Pⱼ) is a homological summand of an abelian carrier
This is a statement about the residual core after retained t₂ factors have been separated; arbitrary surface motives are not declared abelian type.
Back to this node in the diagram ↑Draft · audit pendingu:M=⊕ h(Sα)⊗Lᵗα → h(D), uv=1
Blowup and projective-bundle formulas hold after every smooth projective test product. Original closed product strata, their exact Tate exponents, boundary Chern classes and relative refined intersections are retained.
Back to this node in the diagram ↑Draft · audit pendingA generator with t₂(Xⱼ) factors through fⱼ*
The restricted tower over Uⱼ is Uⱼ×W₋ⱼ. Its actual relative Chow injection agrees there with the exterior product. Localization puts their difference over ∂Xⱼ; a full-degree alteration of each support shows that pⱼ annihilates its cohomology action.
Back to this node in the diagram ↑Draft · audit pendingQ=Q u_ab pr_ab v, with actual correspondence maps
The refined surface complement is Artin–Tate/Picard/Albanese; curve and point factors are abelian type. The resulting u_Qv_Q=Q is a homological retraction, without requiring a new Chow-idempotent lift.
Back to this node in the diagram ↑LOGIC 04
From the model sectors to an algebraic inner inverse
One common coefficient prime compares the abelian carriers with numerical motives; only their operator is inverted, then the retained transcendental factors are tensored back.
Scope. All Chow groups have rational coefficients. Vʳ means geometric homological triviality, component by component. Local assertions use the original finite valued field K/ℚₚ; global assertions use a number field F, n=dim X and 1≤r≤n, s=n+1−r.
Scroll inside the frame to follow a large branch. “Fit width” shows the whole layout; “Readable size” restores the text to 100%.
Read the 2 connections
Each numbered card gives the exact inputs, the deduction and its sources.
01AND · Sector decomposition + carrier inverse + tensoring and finite descent
Premises and stated hypotheses
- Draft · audit pendingThe orthogonal sectors exhaust the actual component cohomology
R_I=P_I∏ₖ∉I(1−Pₖ), Σ_I R_I=1; on each sector A is the identity on the retained t₂ tensor times the residual abelian-carrier operator.
Scope & sources - Published resultChoose one prime for all powers of one common abelian carrier
The finite list of carriers is included in powers and Tate shifts of one B. The stated Deligne–Clozel comparison gives one invertible ℓ where numerical and geometric homological equivalence agree on all its powers; finite-constant norms retain native degrees.
Scope & sources - Draft · audit pendingInvert the operator in the numerical category
Jannsen semisimplicity gives aga=a in the opposite rectangular corner of the source and target summands. Choose an actual rational Chow representative g, then transport G_Q=u_Qgv_Q. The selected-prime comparison gives the identity in cohomology.
Scope & sources - Draft · audit pendingTensor the core inverses, transport by f_I and sum
Retain every degree and Tate label. Tensor each inverse with the identity of its fixed t₂ factors and project by R_I. The empty-factor core has A=0. Algebraic curve/surface degree projectors come from the explicit generator decomposition.
Scope & sources - Draft · audit pendingDescend finite-field choices and add the good coefficient
Normalized Galois averaging preserves the identity because it is linear in G and A descends. For smooth projective T/R use Δ(T₀)⊠A and Δ(T₀)⊠G on every Künneth summand, including mixed classes.
Scope & sources
How the deduction works
The identity is required on full geometric cohomology, rather than only its algebraic subspace. This makes the later correction apply to all mixed classes.
Conclusion / target
A is the actual restriction–Gysin incidence of the ordered model. The all-factor proof supplies one suitable invertible prime, not an all-prime assertion.
Scope & sourcesEvidence and further reading
Back to connection 01 in the diagram ↑All 9 nodes: scope, status and sources
Draft · audit pendingAn actual G satisfies AᴴGᴴAᴴ=Aᴴ on all degrees
A is the actual restriction–Gysin incidence of the ordered model. The all-factor proof supplies one suitable invertible prime, not an all-prime assertion.
Back to this node in the diagram ↑Draft · audit pendingThe orthogonal sectors exhaust the actual component cohomology
R_I=P_I∏ₖ∉I(1−Pₖ), Σ_I R_I=1; on each sector A is the identity on the retained t₂ tensor times the residual abelian-carrier operator.
Back to this node in the diagram ↑Published resultChoose one prime for all powers of one common abelian carrier
The finite list of carriers is included in powers and Tate shifts of one B. The stated Deligne–Clozel comparison gives one invertible ℓ where numerical and geometric homological equivalence agree on all its powers; finite-constant norms retain native degrees.
Back to this node in the diagram ↑Draft · audit pendingInvert the operator in the numerical category
Jannsen semisimplicity gives aga=a in the opposite rectangular corner of the source and target summands. Choose an actual rational Chow representative g, then transport G_Q=u_Qgv_Q. The selected-prime comparison gives the identity in cohomology.
Back to this node in the diagram ↑Draft · audit pendingTensor the core inverses, transport by f_I and sum
Retain every degree and Tate label. Tensor each inverse with the identity of its fixed t₂ factors and project by R_I. The empty-factor core has A=0. Algebraic curve/surface degree projectors come from the explicit generator decomposition.
Back to this node in the diagram ↑Draft · audit pendingDescend finite-field choices and add the good coefficient
Normalized Galois averaging preserves the identity because it is linear in G and A descends. For smooth projective T/R use Δ(T₀)⊠A and Δ(T₀)⊠G on every Künneth summand, including mixed classes.
Back to this node in the diagram ↑Draft · audit pendingFor one surface: the inverse works on full component cohomology
On surface components A has codimension 3 and its numerical generalized inverse G has codimension 1. The error AGA−A is a numerically zero one-cycle on each fourfold Dᵢ×Dⱼ.
Back to this node in the diagram ↑Draft · audit pendingChoose an actual divisor representative of the numerical inverse
Split source kernel and target image in the semisimple numerical category, invert the remaining isomorphism, and lift its finitely many entries to rational divisor correspondences.
Back to this node in the diagram ↑Draft · audit pendingThe numerical-zero error one-cycle has zero ℓ-adic class
Use the algebraic inverse of L²:H²→H⁶ on a product of surfaces. It sends the error to a numerical-zero divisor, hence a homological-zero divisor; apply L² again. Native-to-geometric numerical zero uses finite norms and the full degree.
Back to this node in the diagram ↑LOGIC 05
Chow motives generated by curves and surfaces
Actual inclusions and retractions transfer all shifted classes; arbitrary abelian factors enter through a generically finite addition map from a product of curves.
Scope. All Chow groups have rational coefficients. Vʳ means geometric homological triviality, component by component. Local assertions use the original finite valued field K/ℚₚ; global assertions use a number field F, n=dim X and 1≤r≤n, s=n+1−r.
Scroll inside the frame to follow a large branch. “Fit width” shows the whole layout; “Readable size” restores the text to 100%.
Read the 2 connections
Each numbered card gives the exact inputs, the deduction and its sources.
01AND · Dominant addition + equal dimension + the degree projection formula
Premises and stated hypotheses
- Internal proof availableThe addition map Cᵍ→A is dominant of degree d
Source and target both have dimension g. Dimension one is immediate and dimension zero is the unit motive.
Scope & sources - Internal proof availablef∗f*=dΔ_A gives the retraction d⁻¹f∗
This is a Chow correspondence identity after every smooth projective test product.
Scope & sources
How the deduction works
Successive sums of C grow until they fill A: a proper stabilized image would be invariant under the group generated by C. Therefore Cᵍ→A is generically finite of degree d>0.
Conclusion / target
After finite extension choose a pointed smooth complete-intersection curve C with C−C generating A.
Scope & sourcesEvidence and further reading
Back to connection 01 in the diagram ↑02AND · An actual diagonal certificate + admission of every intermediate degree
Premises and stated hypotheses
- HypothesisThe inclusion/retraction composites sum to Δ_X
The decomposition is in rational Chow motives with the stated Tate shifts. Every image class remains geometrically homologically trivial.
Scope & sources - Internal proof availableA g-dimensional abelian factor retracts from Cᵍ
After finite extension choose a pointed smooth complete-intersection curve C with C−C generating A.
Scope & sources - ConditionalThe conditional curve–surface product theorem
Use the low-dimensional endpoint, empty-product good-reduction, or higher-dimensional product-model branch in its stated range. Retain its analytic hypotheses and, for the last branch, the additional model hypothesis. At every finite place T must have potentially good projective reduction.
Scope & sources
How the deduction works
Insert each inclusion and retraction, tensor with Δ_T when needed, and use the general Chow transfer. Field descent uses the complete degree.
Conclusion / target
After finite extension, h(X) is a summand of a finite sum of Tate twists of curve/surface products. This also covers T×X with the permitted coefficient, and arbitrary abelian factors.
Scope & sourcesEvidence and further reading
Back to connection 02 in the diagram ↑All 6 nodes: scope, status and sources
ConditionalThe full pairing passes to the stated surface-generated Chow domain
After finite extension, h(X) is a summand of a finite sum of Tate twists of curve/surface products. This also covers T×X with the permitted coefficient, and arbitrary abelian factors.
Back to this node in the diagram ↑HypothesisThe inclusion/retraction composites sum to Δ_X
The decomposition is in rational Chow motives with the stated Tate shifts. Every image class remains geometrically homologically trivial.
Back to this node in the diagram ↑Internal proof availableA g-dimensional abelian factor retracts from Cᵍ
After finite extension choose a pointed smooth complete-intersection curve C with C−C generating A.
Back to this node in the diagram ↑Internal proof availableThe addition map Cᵍ→A is dominant of degree d
Source and target both have dimension g. Dimension one is immediate and dimension zero is the unit motive.
Back to this node in the diagram ↑Internal proof availablef∗f*=dΔ_A gives the retraction d⁻¹f∗
This is a Chow correspondence identity after every smooth projective test product.
Back to this node in the diagram ↑ConditionalThe conditional curve–surface product theorem
Use the low-dimensional endpoint, empty-product good-reduction, or higher-dimensional product-model branch in its stated range. Retain its analytic hypotheses and, for the last branch, the additional model hypothesis. At every finite place T must have potentially good projective reduction.
Back to this node in the diagram ↑LOGIC 06
Hilbert schemes of points on surfaces
The incidence correspondences reduce each homological class, with its exact dimension shift, to powers of the original surfaces.
Scope. All Chow groups have rational coefficients. Vʳ means geometric homological triviality, component by component. Local assertions use the original finite valued field K/ℚₚ; global assertions use a number field F, n=dim X and 1≤r≤n, s=n+1−r.
Scroll inside the frame to follow a large branch. “Fit width” shows the whole layout; “Readable size” restores the text to 100%.
Read the 1 connection
Each numbered card gives the exact inputs, the deduction and its sources.
01AND · The actual Hilbert-scheme motive decomposition + the product theorem
Premises and stated hypotheses
- Published resulth(S^[m])=⊕λ (h(Sˡ⁽λ⁾),eλ)⊗Lᵐ⁻ˡ⁽λ⁾
For λ=1ᵃ¹⋯mᵃᵐ, l(λ)=Σaᵢ and eλ is the rational average of the graphs of ∏ᵢSₐᵢ. The stated de Cataldo–Migliorini Theorem 6.2.1 and Remark 6.2.2 give actual incidence correspondences.
Scope & sources - Internal proof availableA codimension-r class maps to codimension r−s
For a tuple of partitions, s=Σ(mᵢ−l(λᵢ)); the intermediate variety is T×∏Cⱼ×∏Sᵢˡ⁽λᵢ⁾ and has dimension dim X−2s. Images remain homological; degree-zero rank kernels and impossible degrees vanish.
Scope & sources - ConditionalThe conditional curve–surface product theorem
Use the low-dimensional endpoint, empty-product good-reduction, or higher-dimensional product-model branch in its stated range. Retain its analytic hypotheses and, for the last branch, the additional model hypothesis. At every finite place T must have potentially good projective reduction.
Scope & sources - ConditionalThe incidence composites recover the actual diagonal
The admitted intermediate classes return to the original class through Chow covariance. Apply the two-degree criterion at every completion.
Scope & sources
How the deduction works
Tensor the incidence correspondences and their projectors, then transfer every feasible degree. The same correspondences base change to treat every new class after finite extension.
Conclusion / target
The surfaces are arbitrary smooth projective geometrically integral surfaces and mᵢ≥1. T has potentially good projective reduction at every finite place.
Scope & sourcesEvidence and further reading
Back to connection 01 in the diagram ↑All 5 nodes: scope, status and sources
ConditionalAll pairings on T × ∏Cⱼ × ∏Sᵢ^[mᵢ]
The surfaces are arbitrary smooth projective geometrically integral surfaces and mᵢ≥1. T has potentially good projective reduction at every finite place.
Back to this node in the diagram ↑Published resulth(S^[m])=⊕λ (h(Sˡ⁽λ⁾),eλ)⊗Lᵐ⁻ˡ⁽λ⁾
For λ=1ᵃ¹⋯mᵃᵐ, l(λ)=Σaᵢ and eλ is the rational average of the graphs of ∏ᵢSₐᵢ. The stated de Cataldo–Migliorini Theorem 6.2.1 and Remark 6.2.2 give actual incidence correspondences.
Back to this node in the diagram ↑Internal proof availableA codimension-r class maps to codimension r−s
For a tuple of partitions, s=Σ(mᵢ−l(λᵢ)); the intermediate variety is T×∏Cⱼ×∏Sᵢˡ⁽λᵢ⁾ and has dimension dim X−2s. Images remain homological; degree-zero rank kernels and impossible degrees vanish.
Back to this node in the diagram ↑ConditionalThe conditional curve–surface product theorem
Use the low-dimensional endpoint, empty-product good-reduction, or higher-dimensional product-model branch in its stated range. Retain its analytic hypotheses and, for the last branch, the additional model hypothesis. At every finite place T must have potentially good projective reduction.
Back to this node in the diagram ↑ConditionalThe incidence composites recover the actual diagonal
The admitted intermediate classes return to the original class through Chow covariance. Apply the two-degree criterion at every completion.
Back to this node in the diagram ↑LOGIC 07
Moduli spaces on K3 and abelian surfaces
A complex Chow retraction is spread together with its rational-equivalence certificate and specialized to a finite number field before local transfer.
Scope. All Chow groups have rational coefficients. Vʳ means geometric homological triviality, component by component. Local assertions use the original finite valued field K/ℚₚ; global assertions use a number field F, n=dim X and 1≤r≤n, s=n+1−r.
Scroll inside the frame to follow a large branch. “Fit width” shows the whole layout; “Readable size” restores the text to 100%.
Read the 2 connections
Each numbered card gives the exact inputs, the deduction and its sources.
01AND · Spread the actual equations and choose a closed point of a smooth base
Premises and stated hypotheses
- Internal proof availableSpread correspondences and their finite principal-divisor witnesses
Descend the data to a finitely generated field, spread over a smooth integral finite-type F-scheme B, and shrink so fibers are smooth projective geometrically integral and cycle codimensions and the relative Chow equality are preserved.
Scope & sources - Internal proof availableA closed point b has finite residue field E/F
Since B is smooth, the refined fiber pullback preserves the finite certificate. This is specialization of the actual Chow equality, not just of its realization.
Scope & sources
How the deduction works
Refined pullback at that point commutes with smooth pullback, proper projection and diagonal intersection, hence with the composition in the certificate.
Conclusion / target
The constant X specializes to X_E, while the finitely many auxiliary families specialize to smooth projective curves and surfaces with all degree shifts preserved.
Scope & sourcesEvidence and further reading
Back to connection 01 in the diagram ↑02AND · The precise complex retraction + finite specialization + local transfer
Premises and stated hypotheses
- Published resultBülles’s Theorem 0.1 supplies a rational Chow retraction
The intermediate motives are Tate twists of powers of the underlying surface. Keep smoothness, projectivity and the stated moduli/stability hypotheses; finite-dimensionality of the surface motive is not required.
Scope & sources - Internal proof availableA full Chow certificate specializes to finite E/F
The constant X specializes to X_E, while the finitely many auxiliary families specialize to smooth projective curves and surfaces with all degree shifts preserved.
Scope & sources - ConditionalComplete E/F and transfer at every original finite place
Use the surface-generated domain at a place above v, then full-degree native descent to F_v. For new cycles after finite extension repeat over a compositum and apply both-degree admission.
Scope & sources - ConditionalProducts and Hilbert-scheme factors use tensor certificates
Intermediate varieties remain curve/surface products; the optional T is allowed when it has potentially good projective reduction at every finite place.
Scope & sources
How the deduction works
The first theorem also applies to any X whose complex Chow motive retracts through products of curves and surfaces. The auxiliary varieties may change under specialization.
Conclusion / target
For a smooth projective X/F, one complex embedding must identify its motive with the retraction supplied by Bülles: Gieseker-stable twisted sheaves, or stable objects for a generic stability condition, on a projective K3 or abelian surface.
Scope & sourcesEvidence and further reading
Back to connection 02 in the diagram ↑All 7 nodes: scope, status and sources
ConditionalThe stated smooth projective moduli spaces have every pairing
For a smooth projective X/F, one complex embedding must identify its motive with the retraction supplied by Bülles: Gieseker-stable twisted sheaves, or stable objects for a generic stability condition, on a projective K3 or abelian surface.
Back to this node in the diagram ↑Published resultBülles’s Theorem 0.1 supplies a rational Chow retraction
The intermediate motives are Tate twists of powers of the underlying surface. Keep smoothness, projectivity and the stated moduli/stability hypotheses; finite-dimensionality of the surface motive is not required.
Back to this node in the diagram ↑Internal proof availableA full Chow certificate specializes to finite E/F
The constant X specializes to X_E, while the finitely many auxiliary families specialize to smooth projective curves and surfaces with all degree shifts preserved.
Back to this node in the diagram ↑Internal proof availableSpread correspondences and their finite principal-divisor witnesses
Descend the data to a finitely generated field, spread over a smooth integral finite-type F-scheme B, and shrink so fibers are smooth projective geometrically integral and cycle codimensions and the relative Chow equality are preserved.
Back to this node in the diagram ↑Internal proof availableA closed point b has finite residue field E/F
Since B is smooth, the refined fiber pullback preserves the finite certificate. This is specialization of the actual Chow equality, not just of its realization.
Back to this node in the diagram ↑ConditionalComplete E/F and transfer at every original finite place
Use the surface-generated domain at a place above v, then full-degree native descent to F_v. For new cycles after finite extension repeat over a compositum and apply both-degree admission.
Back to this node in the diagram ↑ConditionalProducts and Hilbert-scheme factors use tensor certificates
Intermediate varieties remain curve/surface products; the optional T is allowed when it has potentially good projective reduction at every finite place.
Back to this node in the diagram ↑Remaining problem
The additional step is the product-model calculation on full cohomology, including its transcendental summands and exceptional components. This argument does not reduce an arbitrary surface motive to an abelian-type motive.