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.
Finite-parameter smoothing of vector-bundle sections
Statement
Let be a compact finite-dimensional smooth manifold, a closed smooth submanifold with a chosen collar, a compact smooth manifold, and a finite-rank smooth real vector bundle. Let be a continuous family of sections, meaning and is continuous. Assume is continuous in the uniform topology, its restriction to is continuous in the section topology, and for distinct marked points the values are independent of . For every there is a family continuous in the topology, equal to on , equal to at every , and satisfying for any fixed bundle norm. The straight-line homotopy in each vector fiber from to fixes and all marked values.
Statement references
Facts & Assumptions
Given: The compact manifolds, vector bundle, collar, family, fixed fiber values and uniform bundle norm of the Statement. Work in ZF + AC to use the supplied smooth partition-of-unity theorem.
A smooth vector bundle is locally a product with finite-dimensional vector fibers, and its smooth transition functions identify the local section coefficients. Vector bundle charts and transition functions
On a smooth manifold, a smooth partition of unity subordinate to an open cover supplies smooth cutoff functions supported in the chosen charts. Smooth partitions of unity exist on manifolds, with the convention of Smooth partitions of unity subordinate to an open cover
The family on is continuous into the full topology, not merely pointwise smooth in ; this is the relative-topology hypothesis used below.
Proof
If is empty, take the unique empty family. If is empty, its bundle has the unique empty smooth section, and take the unique family of those sections. In either case all agreement constraints and the straight-line homotopy hold vacuously, and the displayed error is . We may therefore assume and nonempty. Choose a finite bundle atlas and a subordinate finite smooth partition of unity using compactness and [F1, F2]. In each chart multiply a section coefficient by its partition function, extend by zero within the chart, convolve with a fixed compactly supported smooth mollifier of radius , and sum the finitely many resulting local sections. For each this defines a linear operator : every output derivative is an integral of the input against a derivative of a fixed smooth kernel, so each seminorm is bounded by a finite constant times the input norm. Uniform continuity of each input on compact chart supports gives for every . Since has compact image in , a finite subcover argument makes this convergence uniform over .
The family is continuous into by the seminorm bounds in step 1.1. Let be the chosen collar neighborhood of , with projection , and choose a continuous cutoff equal to on and supported inside . For an interior hypersurface use the two-sided collar , , with supported in ; no boundary of is needed. Set on and off . Every term is a smooth section in and varies continuously in with . At , . As approaches , continuity into and the uniform approximation in step 1.1 make the correction uniformly small; choose the collar width and then so remains arbitrarily -close to .
For each marked point choose a bundle chart and a smooth bump supported in that chart with ; shrink the charts to be pairwise disjoint. In the chart extend as the constant coefficient vector and multiply by . The resulting correction is a smooth section continuous in in every seminorm, is zero on because there, and changes the value at by exactly . Put . The finite corrections preserve continuity and the relative condition. Since the errors at the marked points are bounded by the uniform error from step 2.1 and the bumps have fixed bounded norms, choosing that error sufficiently small gives the prescribed bound.
For define in the vector space . This is continuous, equals at and at , and is constant on and at each marked point because both endpoint families agree there. Thus the required relative homotopy exists.