Alphabeta Math
LemmaStatement: AI-adaptedProof: AI-generated
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.

Reduced bundle class identities

PDF p. 11: Reduced rank-two identity; PDF p. 12: Induction identity

Statement

Assume the Axiom of Choice and the site and universe bounds required by the rank-fiber, determinant-fiber and Chern-class comparisons. Let k be any field and X a quasi-compact quasi-separated k-scheme. For a vector bundle V of constant rank n≥0, put δ(V)=[V]−[det⁡V]−(n−1)1∈K0(X). Then δ(V)∈Fmot2K0(X). If V and W have the same constant rank and isomorphic determinant line bundles, then [V]−[W]=δ(V)−δ(W). If V=L1⊕L2 for line bundles L1,L2, then ([L1]−1)([L2]−1)=[det⁡V]+1−[V]. For every V as above, its virtual second Chern class satisfies c2(δ(V))=c2(V). Finally, if L is a line bundle, then δ(V⊕L)=δ(V)−([L]−1)([det⁡V]−1).

A vector bundle is a finite locally free OX-module; a line bundle has rank one. The determinant det⁡V is ⋀nV, with det⁡0=OX. Quasi-compact means that every open cover has a finite subcover; quasi-separated means that intersections of quasi-compact opens are quasi-compact. In particular, these assertions apply to all normal qcqs schemes over an algebraically closed field, including the finite-type case. Normal means that every local ring is an integrally closed domain. Here K0(X) is the zeroth group of nonconnective K-theory of perfect complexes; a vector bundle gives a class in it, 1=[OX], and its product is induced by tensor product. A perfect complex is locally quasi-isomorphic to a bounded complex of vector bundles. The group FmotjK0(X) is the image of the jth motivic filtration stage in K0(X). The virtual Chern class in the displayed formula is the second coefficient of the representative-independent total Chern series on the vector-bundle Grothendieck group, evaluated on the indicated virtual bundle expression; no identification of that group with all of perfect-complex K0(X) is assumed. Choice asserts that every family of nonempty sets admits a choice function; the size bounds fix the sites and ambient universes of the cited comparisons.

Facts & Assumptions

Given: k, X, Choice, the stated size bounds, and the vector bundles and line bundles occurring in the assertions.

Goal: Prove the second-filtration membership, both ring identities, the equal-rank comparison, and the virtual second-Chern identity.

[F1]

The zeroth filtration stage is K(X) and the image groups are the images of the stage augmentations. On qcqs field-schemes, put S=Ω∞Fmot1K and T=Ω∞Fmot2K. The actual first projection d:S→BGm is determinant, and T is its homotopy fiber over the trivial line. Thus a first-stage point together with a trivialization of its determinant gives a second-stage lift. Motivic filtration and image groups The actual determinant fiber is a natural retract of the rank-zero K-space

[F2]

For every qcqs scheme over a field, the actual first-stage augmentation identifies π0Fmot1K(X) injectively with the kernel of Euler rank on perfect K0(X). Hence every rank-zero perfect class has a unique first-stage class lifting it. The actual first EM stage is the rank-zero K space on all qcqs schemes

[F3]

The first perfect Chern class of a virtual bundle is the image of its determinant under the weight-one orientation; this agrees with the first coefficient of its virtual bundle Chern series. First Chern class factors through determinant

[F4]

The total Chern series is multiplicative for direct sums and virtual differences. For a line bundle M, ct(M)=1+c1(M)t, and c1(V)=c1(det⁡V). The latter follows from [F3], while the line and virtual-series formulas are supplied by Whitney identities and virtual Chern arithmetic.

[F5]

The Axiom of Choice is among the hypotheses of the cited filtration and Chern comparisons. The Axiom of Choice

Proof

technique · Lift the rank-zero class through the actual determinant fiber, then compute in the bundle Grothendieck ring and its virtual Chern series
1.1givenF1F2F5

The rank of δ(V) is n−1−(n−1)=0, including n=0. By [F2], it has a unique first-stage class s∈π0S(X) lifting it. Its virtual determinant is det⁡V⊗(det⁡V)−1⊗(det⁡OX)−(n−1)≅OX, since the determinant of a virtual trivial bundle is trivial. By [F1], the image of s under the actual determinant map is therefore the trivial line class. Represent s by a point of S(X) and choose a trivialization of its determinant; this is a path of its image to the basepoint of BGm(X). The point and path define a point of the actual fiber T(X) whose underlying perfect class is δ(V). By the image-group definition in [F1], this proves δ(V)∈Fmot2K0(X). These suppliers apply throughout the stated qcqs scope under the assumed Choice and size bounds.

1.2givenconstruct

If V and W have the same rank and isomorphic determinants, subtracting their definitions of δ cancels the determinant and trivial-bundle terms. This gives δ(V)−δ(W)=[V]−[W].

1.3givenconstruct

For V=L1⊕L2, its determinant is L1⊗L2. Distributivity in K0(X) and the direct-sum relation give ([L1]−1)([L2]−1)=[L1⊗L2]−[L1]−[L2]+1=[det⁡V]+1−[V]. This calculation is performed on actual bundle classes and their images in perfect-complex K0.

1.4F3F4construct

Write a=c1(V)=c1(det⁡V). By [F4], the total series of δ(V) is ct(V)(1+at)−1: the trivial-bundle factor has total class 1, even when its coefficient in the virtual expression is negative. Through degree two, ct(V)(1+at)−1=(1+at+c2(V)t2+⋯ )(1−at+a2t2+⋯ ). The coefficient of t2 is c2(V)−a2+a2=c2(V). This proves the stated virtual identity without requiring a Chern-series operation on arbitrary perfect classes.

2.1givenconstructstep 1.1step 1.2step 1.3step 1.4∎

For D=det⁡V, the determinant of V⊕L is D⊗L. Therefore δ(V⊕L)−δ(V)=[L]−[D⊗L]+[D]−1=−([L]−1)([D]−1). The equality uses distributivity and the tensor-product interpretation of multiplication. Together with steps 1.1 through 1.4 this proves all assertions.

Depends on

Used by · 1

Dependency tree · 3 levels

Arrows run from prerequisites to the items that use them. External source nodes link directly to the cited text. The view shows up to three dependency steps; open a prerequisite to continue.

Sources