Alphabeta Math
LemmaStatement: AI-adaptedProof: AI-adaptedPipeline-generatedprecheck passjudge pass (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.

Formal immersion gives the tangent normal-bundle identity

Statement

Assume ACω. For every formal immersion (f,F) from Mm to Nn the image F(TM) is a smooth subbundle of the pullback f∗TN, and the quotient normal bundle νF:=f∗TN/F(TM) has rank n−m when m≤n; if M=∅ and m>n, use the empty rank-zero bundle as in Normal bundle of a formal immersion. The quotient map exhibits the short exact sequence of smooth vector bundles over M 0⟶TM→ F f∗TN⟶νF⟶0, which splits over M: a smooth complement C⊆f∗TN of F(TM) restricts to an isomorphism C→νF and gives a smooth bundle isomorphism TM⊕νF→f∗TN, (v,q(c))↦F(v)+c. The splitting is not canonical in general; if a bundle metric on f∗TN is chosen, the orthogonal complement F(TM)⊥ is a canonical complement for that metric, the orthogonal splitting TM⊕F(TM)⊥≅f∗TN restricts to F on the tangent summand, and different metrics give isomorphic splittings. In particular, for a genuine immersion the isomorphism identifies the normal bundle of the immersion with the quotient f∗TN/df(TM).

Facts & Assumptions

Given: ACω and smooth manifolds Mm,Nn and a formal immersion (f,F) with F:TM→TN a smooth bundle map over f that is injective on every fibre.

[F1]

The pullback f∗TN is a smooth vector bundle over M, and a smooth bundle map over f is the same as a smooth section of Hom⁡(TM,f∗TN) (The pullback fibre product is a smooth vector bundle, Bundle maps over f are sections of the pulled-back Hom bundle, Pullback vector bundles as fibre products).

[L1]

The quotient of a smooth vector bundle by a smooth subbundle is a smooth vector bundle, with the quotient bundle map over the identity (A vector bundle quotient by a subbundle is a smooth vector bundle, Quotient vector bundles by a subbundle).

[L2]

Under ACω, every smooth subbundle of a smooth vector bundle has a smooth complement, and a smooth bundle metric produces the orthogonal complement as a smooth subbundle (Every vector subbundle has a smooth complement, Orthogonal complements of subbundles are smooth subbundles, Every smooth vector bundle admits a smooth bundle metric).

[L3]

The normal bundle of an embedded submanifold is the quotient of the ambient tangent bundle restricted to it by the tangent bundle, and an ambient Riemannian metric identifies it with the orthogonal normal bundle (Normal and conormal bundles of an embedded submanifold, Assuming countable choice, an ambient metric identifies the two normal bundles).

Proof

technique · direct
1.1F1givenconstruct

If M=∅, all total spaces and maps in the claimed sequence and splitting are empty, so exactness and the isomorphisms hold vacuously, with the stated normal-rank convention. Hence assume M is nonempty, so m≤n, and fix x0∈M. In a chart U of M around x0, a trivialization of TM∣U and a trivialization of TN over a chart containing f(U), the section of [F1] is given by a smooth matrix function A of rank m at every point. Reordering coordinates we may suppose an m×m block B(x) of A(x) is invertible near x0; then the image of A(x) equals the image of the block matrix (ImC(x)) with C=A2B−1 smooth, so the image is a smooth subbundle over U with the columns of (IC) as a smooth frame.

2.1F1L1step 1.1

The local frames of step 1.1 agree on overlaps because they span the same subspace at every point, so they glue to a smooth subbundle F(TM)⊆f∗TN of rank m. The quotient νF:=f∗TN/F(TM) is a smooth vector bundle by [L1], and the quotient map is a smooth bundle map over id⁡M; composing the fibrewise isomorphisms Fx:TxM→F(TM)x with the inclusion gives a smooth bundle map TM→f∗TN with image F(TM) and kernel 0, so 0→TM→Ff∗TN→νF→0 is a short exact sequence of smooth vector bundles; ranks give rank⁡νF=n−m.

3.1F1L2step 2.1algebra

Let C⊆f∗TN be a smooth complement of F(TM), which exists by [L2]. Fibrewise, Fx⊕id⁡:TxM⊕Cx→f∗TNx is injective between spaces of dimension n, hence an isomorphism; it is smooth as a bundle map over id⁡M, so it is a smooth bundle isomorphism TM⊕C→f∗TN restricting to F on TM. The quotient map restricts to an isomorphism C→νF because C⊕F(TM)=f∗TN and C∩F(TM)=0; composing the inverse of this isomorphism with the isomorphism above gives TM⊕νF≅f∗TN, (v,q(c))↦F(v)+c. The construction depends on the choice of C; the formula displays that dependence, and no complement is distinguished without further data, so the splitting is not canonical.

4.1L2step 3.1

Choose a smooth bundle metric on f∗TN, which exists by [L2]. Its orthogonal complement F(TM)⊥ is a smooth subbundle, is a complement of F(TM), and is canonically determined by the metric; step 3.1 applied to it gives the orthogonal splitting, and applying step 3.1 to two different complements C,C′ exhibits both νF-decompositions as isomorphic, since both are identified with the quotient.

5.1L3step 2.1step 4.1∎

For a genuine immersion (f,F)=(f,df) the image is df(TM) and the same sequence exhibits νdf=f∗TN/df(TM); when f is an embedding this is the normal bundle of the embedded image by [L3], where the metric identification with the orthogonal normal bundle is precisely the construction of step 4.1.

Depends on

Used by

Cited to discharge well-definedness by Normal bundle of a formal immersion.

Dependency tree · two levels

48 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