Alphabeta Math
LemmaStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedprecheck passverified 2026-11-04 (gpt-6.1-sol)
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

Wu formula for a closed smooth manifold

Statement

Assume full AC. Let M be a connected compact smooth n-manifold without boundary. Let vk(M) be the Wu classes defined by the mod-two Poincare pairing and let wi(TM) be the Stiefel–Whitney classes of the real tangent bundle. For the total classes v=∑k=0nvk and w=∑i=0nwi, with Sq=∑j≥0Sqj, the identity is w=Sq(v)in H∗(M;F2).

Facts & Assumptions

Given: Full AC and a connected closed smooth n-manifold M, with its canonical mod-two fundamental class [M]2.

[A1]

Full AC is The Axiom of Choice, as inherited from the Wu-class and Poincare-duality suppliers.

[F1]

The Wu classes are characterized by ⟨vkx,[M]2⟩=⟨Sqkx,[M]2⟩ for x∈Hn−k(M;F2); the mod-two cup pairing is nonsingular (Wu classes of a closed manifold, Poincaré duality gives a nonsingular cup pairing).

[F2]

The canonical mod-two orientation makes M an F2-oriented closed manifold. The perfect mod-two Poincare pairing supplied by Poincaré duality gives a nonsingular cup pairing also gives finite-dimensional cohomology groups, so homogeneous bases and their Poincare-dual bases exist.

[F3]

The diagonal Δ:M↪M×M is a closed smooth embedding whose normal bundle is canonically isomorphic to TM by (u,v)↦v−u. Its mod-two Poincare-dual class uΔ∈Hn(M×M;F2) is the image of its normal Thom class and satisfies uΔ∩[M×M]2=Δ∗[M]2 (The normal Thom class realizes the Poincare dual of a closed submanifold). The Kronecker pairing is compatible with cap-duality, cap-cup products, and pushforward under Δ (The cap-duality map of an oriented manifold, Kronecker evaluation pairing, Cap naturality and projection formula).

[F4]

The cohomological Künneth theorem makes the external product a graded-ring isomorphism when R is a PID and one factor has finite-free homology in every degree (Cohomological Kunneth cross product is a ring isomorphism). Here R=F2 is a field, hence a PID, and the finite-generation part of the Poincaré-duality supplier's proof gives finite-dimensional Hj(M;F2) for every j; finite-dimensional vector spaces are finite free, so the Künneth hypotheses hold. For external products the evaluation formula is ⟨x×y,[M×M]2⟩=⟨x,[M]2⟩⟨y,[M]2⟩: it follows by evaluating external cochains on the singular chain cross product representing [M]2×[M]2, which is the local mod-two fundamental class of M×M. Thus the product cup pairing pairs decomposable classes by the product of the pairings on the factors, with signs trivial over F2 (Poincaré duality gives a nonsingular cup pairing).

[F5]

The Thom identity gives Sqi(uΔ)=wi(νΔ)⌣uΔ, and the Whitney formula identifies the characteristic classes of the diagonal normal bundle with those of TM (Thom identity for Stiefel–Whitney classes, Whitney sum formula for Stiefel–Whitney classes, Stiefel–Whitney classes from the projective-bundle relation).

[F6]

Steenrod squares are natural and satisfy the Cartan formula for external products; Sq0 is the identity and positive squares of the degree-zero unit vanish (Cartan formula for Steenrod squares, Steenrod normalization, instability, suspension, and top square).

Proof

technique · identify the diagonal class by dual bases, then apply the Thom identity and Cartan formula
1.1A1F1F2choose

If M=∅, its cohomology groups are zero, so w=Sq(v)=0 and the identity holds. Henceforth suppose M is nonempty. For each d, choose a basis {bd,r}r of Hd(M;F2) and its Poincare-dual basis {bd,r}r⊆Hn−d(M;F2), so ⟨bd,r⌣be,s,[M]2⟩=δdeδrs. The finite-dimensionality in [F2] makes these finite bases, and nonsingularity gives the dual bases.

2.1F1F3F4step 1.1algebra

The Künneth isomorphism in [F4] and diagonal identity in [F3] give uΔ=∑d,rbd,r×bd,r. Indeed, for any decomposable a×b of complementary degree, the evaluation formula in [F4] shows that pairing the right side with it gives ⟨a⌣b,[M]2⟩: the dual-basis identities contract the two finite sums. This equals ⟨uΔ⌣(a×b),[M×M]2⟩ because uΔ∩[M×M]2=Δ∗[M]2 and cap-product naturality gives Δ∗(a×b)=a⌣b. Decomposables span by [F4], and the product pairing is nonsingular by [F1].

3.1F2F4step 2.1algebra

Define the linear map λ:H∗(M×M;F2)→H∗(M;F2) by λ(a×b)=⟨b,[M]2⟩a and linear extension through the Kunneth isomorphism. This is well-defined by the tensor-product description in [F4]. Applying it to step 2.1 retains only d=n; connectedness makes Hn(M;F2)=F2 with its generator evaluating to 1, whose dual in H0 is 1. Thus λ(uΔ)=1. Also λ((c×1)z)=c⌣λ(z) for every c∈H∗(M;F2) and z∈H∗(M×M;F2), first on decomposables and then by linearity.

4.1F3F5step 3.1algebra

The normal bundle identification in [F3] and Thom identity [F5] give Sqi(uΔ)=(wi(TM)×1)⌣uΔ. Applying λ and using step 3.1 yields λ(Sqi(uΔ))=wi(TM). Summing the finite nonzero components gives λ(Sq(uΔ))=w(TM).

4.2F1F4F6step 1.1step 3.1algebra

By Cartan and naturality in [F6], Sq(uΔ)=∑d,rSq(bd,r)×Sq(bd,r). Therefore step 3.1 gives λ(Sq(uΔ))=∑d,r⟨Sq(bd,r),[M]2⟩ Sq(bd,r). The dual-basis expansion and the defining Wu pairing in [F1] say v=∑d,r⟨Sq(bd,r),[M]2⟩bd,r: for each homogeneous degree, pairing this expression with any basis element of complementary degree recovers exactly the defining Steenrod-square functional. Hence the preceding sum is Sq(v).

5.1F1F3F5F6step 4.1step 4.2∎

Steps 4.1 and 4.2 identify the same class λ(Sq(uΔ)) as w(TM) and Sq(v), respectively. Thus w(TM)=Sq(v) in mod-two cohomology. The empty case was settled in step 1.1, and connectedness excludes more than one nonempty component; degrees above n vanish by the rank convention, all sums are finite, and signs in the diagonal and product formulas disappear over F2. This proves the asserted identity without an orientation choice.

Depends on

Dependency tree · 2 levels

Sources