Alphabeta Math
TheoremStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedprecheck passaudited 2026-09-14
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.

Difference cochains classify homotopies of extensions in the stable stage

Statement

Assume AC. Let (X,A) be a relative CW complex, let n≥1, and let Y be (n−1)-connected and n-simple; when n=1, assume in particular that π1(Y) is abelian. Fix a:A→Y and put Π=πn(Y) with its resulting simple coefficient system.

Let En(a) consist of maps g:Xn∪A→Y extending a whose obstruction cochain is zero, modulo homotopy rel A on Xn∪A. Equivalently, these are the n-stage maps which extend over Xn+1∪A, with an extension chosen only when needed. If En(a) is nonempty, then

Hn(X,A;Π)

acts freely and transitively on it. For g0,g1∈En(a), the displacement is the difference class

[d(g0,g1)]∈Hn(X,A;Π),

and this class is zero if and only if g0 and g1 are homotopic rel A through the n-skeleton. For a finite relative CW pair, only finite choice is used.

Facts & Assumptions

[F1]

Since Y is (n−1)-connected, every two extensions of a are homotopic rel A through Xn−1∪A; for n=1, abelianness makes the conjugation action simple.

[F2]

For a chosen prior-stage homotopy, δd(g0,g1)=θ(g0)−θ(g1) (The primary obstruction class is independent of cellular choices).

[F3]

The difference-cochain construction uses one shifted prism cell for each relative cell and identifies its primary obstruction cochain with the signed difference cochain (Difference cochain between two cellular extensions).

[F4]

For every cellular n-cochain d, the vanishing-obstruction theorem constructs a map fd equal to f on the prior skeleton and having prescribed difference d(f,const⁡,fd)=d, by relative pinch maps and simultaneous choice; it makes no claim that f and fd are homotopic on the n-skeleton (Vanishing of the primary obstruction is equivalent to extension over the next skeleton).

[A1]

AC is used only to choose representatives and fillers for arbitrary cell families (The Axiom of Choice).

Proof

Given: (X,A), Y, a, and [A1] as in the statement.

1.1F1F2

Let g0,g1∈En(a). By [F1], choose a homotopy H rel A between their restrictions to Xn−1∪A. Because θ(g0)=θ(g1)=0, [F2] gives δd(g0,H,g1)=0. Thus the difference cochain defines a class in Hn(X,A;Π).

1.2F3

Changing H or changing either endpoint through a homotopy rel A changes this cocycle by a coboundary: apply the obstruction-class independence theorem to the corresponding boundary map on the product pair. Hence [d(g0,g1)] depends only on the two classes in En(a). Reversing a prism changes its sign, and gluing prisms gives

[d(g0,g2)]=[d(g0,g1)]+[d(g1,g2)].

[F3]

2.1A1F2F3F4step 1.1

Regard the homotopy problem as extension over the product pair in [F3]. Suppose [d(g0,H,g1)]=0, and choose c∈Cn−1(X,A;Π) with d(g0,H,g1)=δc. Keep both endpoint maps fixed. On each relative prism en−1×I, whose dimension is n, insert by the relative pinch construction of [F4] a sphere representative of −c(en−1) into the interior of H, leaving its entire boundary, including the two endpoint faces, fixed. AC makes these simultaneous insertions over arbitrary cells; the CW pushout glues them to a new prior-stage homotopy Hc rel A with the same endpoints. On a boundary prism en×I, the only changed faces are the prisms over the (n−1)-cells of ∂en. Their signed incidence sum, with the interval-last sign in [F3], is −δc(en); the same oriented boundary calculation as [F2] therefore gives d(g0,Hc,g1)=d(g0,H,g1)−δc=0 as a cochain, not just as a class. The zero obstruction on each en×I now supplies a filling extending Hc over the relative n-prisms, while the fixed bottom and top faces remain g0 and g1. The glued fillings are a homotopy g0≃g1 on Xn∪A rel A. Conversely, such a homotopy fills every prism, making its difference cochain zero and hence its class zero.

2.2F2F4step 1.1

Fix g∈En(a) and let z∈Zn(X,A;Π). Since g has zero obstruction as in Step 1.1, apply [F4] to a stationary prior-stage homotopy and prescribe difference cochain z. It produces gz:Xn∪A→Y with d(g,const,gz)=z. By [F2],

θ(gz)=θ(g)−δz=0,

so gz∈En(a). [F2, F4]

3.1step 1.2step 2.1step 2.2

If z and z′ differ by a coboundary, Step 1.2 gives [d(gz,gz′)]=[z′−z]=0, so Step 2.1 makes gz and gz′ equivalent. Thus Step 2.2 defines an action of Hn(X,A;Π) on En(a). The addition formula in Step 1.2 proves the action law.

4.1A1step 2.1step 3.1

For any g0,g1, the class [d(g0,g1)] sends g0 to g1, proving transitivity. If a class fixes g0, its displacement is zero by Step 2.1, proving freeness. AC enters only in [F4] and in simultaneous extension over arbitrary cell families; finite families need only finite choice. Empty cell sets, A=X, and the zero group give the asserted singleton torsors. □

Depends on

Dependency tree · 2 levels

Sources