Alphabeta Math
TheoremStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedaudited 2026-09-22
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.

Real splitting principle with mod-two injective pullback

Statement

Assume AC. Let E→B be a numerable real vector bundle of rank n≥0 over a paracompact Hausdorff CGWH base of CW homotopy type, and let q:Fl⁡(E)→B be its real flag bundle. Then Fl⁡(E) is a base of the same kind, q∗E≅L1⊕⋯⊕Lnover Fl⁡(E), and the pullback q∗:H∗(B;F2)⟶H∗(Fl⁡(E);F2) is injective. Moreover, finitely many numerable real bundles E1,…,Em over B admit a common paracompact Hausdorff CGWH base of CW homotopy type X→B with X=Fl⁡(E1)×B⋯×BFl⁡(Em) over which every q∗Ek splits as a sum of line bundles and whose projection X→B induces an injection on F2-cohomology.

Facts & Assumptions

Given: AC, a numerable real rank-n bundle E→B over a paracompact Hausdorff CGWH base of CW homotopy type, and its flag bundle q:Fl⁡(E)→B.

[F1]

The flag bundle is built as the composite of the projections Xj→Xj−1, each of which is the projective bundle of a numerable real bundle of positive rank rj, and q∗E≅L1⊕⋯⊕Ln (Real flag bundle and Stiefel–Whitney roots).

[F2]

For a numerable real rank-r bundle G with r≥1 over a paracompact Hausdorff CGWH base Y of CW homotopy type, the projective bundle theorem gives H∗(P(G);F2) free over H∗(Y;F2) on 1,xG,…,xGr−1; in particular the projection P(G)→Y induces an injection on F2-cohomology, as the inclusion of the coefficient-of-1 summand (Mod-two real projective bundle theorem).

[F3]

Every intermediate Xj and Fl⁡(E) is paracompact Hausdorff CGWH of CW homotopy type, and a fiber product over B of finitely many flag bundles is a numerable bundle with compact CW fiber over B, hence has the same properties (Compact-fibre bundle totals preserve paracompactness, and CW type under CW-type hypotheses).

[F4]

Pullback is canonically functorial and compatible with direct sums, so the splitting of a pulled-back bundle is the pullback of the splitting (Vector-bundle pullback is canonically functorial, Whitney sum, tensor, dual, Hom, and exterior-power bundles).

[A1]

AC is the Axiom of Choice in the form fixed by The Axiom of Choice.

Proof

1.1F1F2F3

Every stage projection is injective on F2-cohomology. By [F1] the map Xj→Xj−1 is the projection of a projective bundle P(Gj) of a numerable bundle of positive rank over Xj−1, and Xj−1 is admissible by [F3]; [F2] therefore makes the induced map H∗(Xj−1;F2)→H∗(Xj;F2) injective. The composite of finitely many injective maps is injective, so q∗:H∗(B;F2)→H∗(Fl⁡(E);F2) is injective.

1.2F1F3

The splitting over the flag bundle is [F1]'s second clause, q∗E≅L1⊕⋯⊕Ln, with each Lj a numerable line bundle over the admissible space Fl⁡(E). For n=0, Fl⁡(E)=B and the sum is empty, so q=id⁡ and the pullback is the identity, which is injective.

1.3

Finitely many bundles. Let E1,…,Em be numerable real bundles over B, of ranks nk≥0. Define Y0=B and recursively Yk:=Yk−1×BFl⁡(Ek),rk:Yk→Yk−1. We first verify the required base change. For a map h:W→Z and a positive-rank bundle G→Z, there is a fiberwise map Ψ:P(h∗G)⟶W×ZP(G),(w,ℓ⊆Gh(w))⟼(w,ℓ). Over a linear chart G∣U≅U×Rr, both sides have the pulled-back projective chart h−1(U)×RPr−1 and Ψ is the identity in these coordinates. Thus Ψ is a homeomorphism over W, and the same coordinate description identifies γh∗G with the pullback of γG. Choose at each flag stage the pullback of the metric used over Z. Fiberwise orthogonal complement then commutes with pullback, because both subbundles consist of the vectors orthogonal to the same pulled-back line. Induction through the projectivization stages of [F1] therefore gives Fl⁡(h∗Ek)≅W×BFl⁡(Ek), with its tautological lines identified with the pulled-back ones.

Apply this with W=Yk−1 and h:Yk−1→B. It identifies rk:Yk→Yk−1 with the flag projection of h∗Ek, so step 1.1 over the admissible base Yk−1 makes every rk∗ injective. With X=Ym=Fl⁡(E1)×B⋯×BFl⁡(Em) and qX=r1∘⋯∘rm, contravariance gives qX∗=rm∗⋯r1∗, a composite of injective maps. By [F3] every Yk, and in particular X, is admissible. For each k, let qk′′:X→Fl⁡(Ek) be the k-th projection. Then qX=qkqk′′ and qX∗Ek≅qk′′∗qk∗Ek≅qk′′∗(L1(k)⊕⋯⊕Lnk(k)), which by [F4] is a direct sum of the line bundles qk′′∗Lj(k). [F1, F3, F4, step 1.1]

2.1F1F2A1step 1.1step 1.3∎

Boundary cases. If some nk=0 then Ek is the zero bundle, its flag bundle is B and qX∗Ek=0 splits as an empty sum; the corresponding factor contributes no projection and does not affect injectivity. If m=0 then X=B, the empty family of bundles is split vacuously, and qX=id⁡. If B=∅ then all spaces are empty and the cohomology groups are zero, so injectivity is vacuous. AC is used through the metric choices of the flag construction and the projective bundle theorem, as recorded.

Depends on

Dependency tree · 2 levels

Sources