Alphabeta Math
TheoremStatement: AI-adaptedProof: AI-adaptedPipeline-generatedprecheck pass
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.

Integral finite generation of universal real and oriented Thom homology

Statement

Assume AC. For every r≥0 and every i≥0, H_i(MO(r);Z) and H_i(MSO(r);Z) are finitely generated. For positive rank the locally constructed homological Thom comparison is H_i(D(γ_r),S(γ_r);Z)=H_{i−r}(BO(r);O_Z(γ_r)), and likewise for γ_r⁺ with the constant orientation system. The disk/sphere quotient identifies these groups with reduced Thom homology. At rank zero the based quotient is B₊; BO(0)=BSO(0)=* gives S⁰ and is handled separately.

Facts & Assumptions

Given: AC; ranks r≥0; the universal metric bundles γr→BO(r) and γr+→BSO(r) with their disk and sphere bundles; the Shubert CW structures on the base with finitely many cells in each dimension and two lifted cells in the oriented case; and the actual relative singular complexes C∗(D,S;Z).

[F1]

The Schubert CW structures have finitely many cells in each dimension; homotopy invariance of pullback, Gram–Schmidt orthonormalization, the two-open-set Mayer–Vietoris sequence for small chains, the homology long exact sequence and the good-pair quotient theorem provide the local product trivializations and the comparison of the algebraic sum with the actual pair complex (Schubert cells give the stable Grassmannian CW structure, Schubert cells in real and complex Grassmannians, Homotopy invariance of vector-bundle pullback, Gram–Schmidt turns every finite independent list into an orthonormal list with the same successive spans, Short exact chain Mayer–Vietoris sequence, Cover-small chains compute singular homology, The long exact sequence in homology, Good pairs and quotient reduced homology).

[F2]

Consecutive CW skeleta have relative homology free on the cells, the first Serre differential is the cellular boundary with local coefficients, and cellular chains compute homology with local coefficients (Relative homology of consecutive CW skeleta, The first Serre differential is the cellular boundary with local coefficients, Cellular chains compute local homology); compact images have finite cell support without choice (Compact CW images have finite cell support without choice).

[F3]

A filtered complex produces an exact couple, which generates a spectral sequence; these algebraic constructions alone assert no abutment (A filtered complex produces an exact couple, An exact couple generates a spectral sequence). The Serre theorem supplies convergence for its fibration hypotheses only (Homological Serre spectral sequence); convergence for the present relative filtration is proved in step 7.1. The sphere and disk computations give the layer homology (Homology of spheres).

[F5]

Over the Noetherian ring Z, submodules of finitely generated modules are finitely generated, so finite free cellular chains have finitely generated homology (Every principal ideal domain is Noetherian, Finitely generated modules over a left Noetherian ring are Noetherian).

Proof

technique · direct
1.1givenF1F4

H_i(D(γ_r),S(γ_r);Z) ≅ H_{i-r}(BO(r);O_Z(γ_r)). Here is the required local homological argument; this is not an assertion that the published cohomological Thom theorem already states it. Write B=BO(r), B_p=B^(p), D_p=D(γ_r)|B_p and S_p=S(γ_r)|B_p. Set D₋₁=S₋₁=∅ and S=S(γ_r). On the actual relative singular complex C=C_(D,S;Z), define the increasing subcomplexes F_p C=(C_(D_p;Z)+C_(S;Z))/C_(S;Z). A simplex common to D_p and S lands in S_p, so F_p C identifies with C_(D_p,S_p). The graded complex is consequently gr_p C=C_(D_p)/(C_(D_{p-1})+C_(S_p)). The sum in this denominator is not silently identified with the singular chains of its union. The following relative collar argument justifies the needed homology comparison.

2.1step 1.1F1F6

Each B_p is finite compact CW, by the Schubert count below. Over a characteristic disk D^p, contract the disk and apply the published homotopy-invariance-of-pullback theorem to trivialize the pulled-back bundle. Orthonormalize the resulting frame with its pulled-back metric. The Gram–Schmidt formulas subtract the earlier orthogonal projections and divide by strictly positive lengths, so all frame coordinates vary continuously; this uses the inspected Gram–Schmidt supplier and needs no unproved continuity of matrix square roots. The pullback fiber pair is therefore the actual product (D^r,S^{r-1}), with transitions preserving the sphere. The finitely many products D^p×D^r, attached over ∂D^p×D^r to D_{p-1}, give D_p with its actual topology. Indeed the attachment quotient maps bijectively to D_p and is compact, whereas D_p is Hausdorff; thus the map is a homeomorphism. This also supplies all subsequent cellwise continuity checks.

3.1step 2.1F1

Put A_p=D_{p-1}∪S_p. In A_p take the two open neighborhoods U and V as follows. U contains D_{p-1} and, in each attached product, the sphere points with base radial coordinate ||x||>1/2. V is the set of fiber points with ||z||>1/2. Their openness follows from the finite product attachment test, they cover A_p, and V deformation retracts to S_p by fiber radial normalization. U deformation retracts to D_{p-1} by radially moving x to x/||x|| in each base-cell collar while retaining the norm-one coordinate z in its isometric trivialization. On the boundary these formulas are the identity, so differing boundary representations give the same actual point. The same homotopy preserves the sphere subbundle and the fiber norm. It retracts U∩V to the norm->1/2 neighborhood of S_{p-1} in D_{p-1}; fiber normalization then retracts that neighborhood to S_{p-1}. For p=0, U is empty and the same statements hold with D₋₁=S₋₁=∅.

4.1step 3.1F1F2

The exact chain sequence for the algebraic sum C_(D_{p-1})+C_(S_p) has intersection C_(S_{p-1}). Compare it with the published two-open-set small-chain sequence for U,V. The three inclusions from D_{p-1}, S_p and S_{p-1} to U,V and U∩V are homology isomorphisms by these retractions. The long exact sequences and their injectivity/surjectivity chase show that the sum inclusion into C_(U)+C_(V) is a homology isomorphism. The inspected thm-cover-small-singular-chains-compute-singular-homology identifies the latter with C_(A_p) on homology. Hence the natural map from gr_p C to C_*(D_p,A_p) is a homology isomorphism, by the short exact quotient sequences. This establishes the relative use of excision without invoking an absolute fiber lemma as if it already treated pairs.

5.1step 4.1F1F3F6

The pair (D_p,A_p) is good. In each product D^p×D^r, A_p contains its full boundary, namely ∂D^p×D^r ∪ D^p×S^{r-1}. The annulus max(||x||,||z||)>1/2, together with A_p itself, is an open neighborhood of A_p and retracts onto A_p by radial normalization in this maximum norm. Boundary points are fixed, so the formula descends through all attachments and is continuous by the same compact quotient test. The good-pair theorem identifies its relative homology with the reduced homology of D_p/A_p. That quotient is a finite wedge, one sphere S^{p+r} for each p-cell of B: each product ball has its entire boundary collapsed and distinct interiors remain distinct. The published sphere and cellular relative calculations give H_{p+q}(gr_p C)=⊕_{p-cells} Z if q=r, and 0 otherwise.

6.1step 5.1F2F4algebra

Orient each base disk and each local fiber. The generator is the base-first product relative orientation class. Under a change of local isometric fiber frame, its sign changes exactly by that frame change's determinant sign. To check d₁ precisely, use the positive chain-connector formula: represent a layer generator by its product disk relative cycle, take its singular boundary, and project that boundary to the preceding layer. The fiber-boundary term is in S and vanishes. The surviving base-boundary term is transported by the disk trivialization. Projecting to each lower-cell summand commutes with this connector by naturality of the quotient and excision maps. Thus a positively oriented attaching incidence acts on the fiber generator by its path transport; reversing that incidence changes its sign by the base disk orientation. In the lifted cellular coordinates of the published local-coefficient cellular theorem, write an attaching boundary as ∂ẽ=Σ_f f̃ r_fe, r_fe=Σ_g n_feg g ∈ Z[π₁(B)]. Each signed term n_feg in this group-ring incidence acts on the local fiber generator by the orientation character ε(g)∈{1,−1}. Consequently our d₁ coefficient is Σ_g n_feg ε(g), exactly the cellular boundary with coefficients O_Z(γ_r). This follows by the preceding connector calculation term by term and its finite additivity, precisely as in the inspected first-Serre-differential supplier; quotienting by fiber sphere chains changes its generator to the relative orientation generator and kills only the fiber-boundary term. All attaching maps have finite support. This formula does not multiply the ordinary summed integer incidence by a single sign: different incidence paths can have different orientation signs. The published lift-basis invariance makes the identity independent of the temporarily supplied cell lifts. There is only the row q=r, so E² is H_p(B;O_Z(γ_r)) on that row and every later differential vanishes.

7.1step 6.1F2F3algebra

Apply the filtered-complex exact-couple construction to FpC, with FpC=0 for p<0. The filtration is exhaustive on chains and on boundary primitives: their projected compact supports lie in finite base subcomplexes. Fix total degree n and column p. In the exact-couple formula of [F3], Np,n−ps=ker⁡k once s>p, because its incoming D1 term has negative filtration index. Moreover Bs=jker⁡(Hn(FpC)→Hn(Fp+s−1C)). Exhaustivity on primitives implies that the union of these kernels is ker⁡(Hn(FpC)→Hn(C)). Step 6.1 shows that all differentials after E2 vanish. Once Ns is fixed, the canonical maps Ns/Bs→Ns/Bs+1 are therefore isomorphisms, so Bs equals its union. Exactness of the initial couple gives ker⁡k=im⁡j and ker⁡j=im⁡(Hn(Fp−1C)→Hn(FpC)). Quotienting consequently identifies the stable term with FpHn(C)/Fp−1Hn(C), where FpHn(C) is the image in Hn(C). The only possible nonzero quotient is at p=n−r. Starting from F−1Hn=0 and using exhaustivity, this single quotient is the entire homology; when n<r, every quotient is zero. Thus Hn(D,S;Z)≅Hn−r(B;OZ(γr)). This proves the required convergence for this filtration without invoking a general abutment theorem.

8.1step 7.1F1F3F5∎

There are only finitely many Schubert cells in each dimension of BO(r): if d=Σ(a_i−i), every a_i≤i+d. The orientation system has stalk Z, so its cellular chains are finite free in each degree. Published cellular homology and Noetherianity give finite generation of their homology. Finally S(γ_r) is a closed subspace of D(γ_r), with open neighborhood ||v||>1/2 retracting onto S by radial normalization. The published good-pair homology theorem identifies relative disk/sphere homology with reduced Thom homology. The basepoint adds only the finitely generated H₀ summand. This proves the assertion. The same argument works for MSO(r), whose base cells are the two lifted copies of the ordinary Schubert cells. At rank zero, handle the assertion separately: D=B, S=∅ and the based Thom quotient is B₊, whose reduced homology is H_*(B;Z). Here B=BO(0)=BSO(0) is a point, so finite generation is immediate. The positive-rank good-pair theorem is not applied to its empty sphere bundle.

Depends on

Used by

Dependency tree · two levels

113 results within two dependency steps of this one, each drawn at its shortest distance from it. An arrow runs from a result to what uses it, so the chart reads left to right and ends at this result, which carries a heavier outline. Every node is a link to that result. Click elsewhere on the chart to enlarge it.

Sources