Alphabeta Math
PropositionStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedjudge 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.

The projective Nakayama pairing and the symmetric-algebra specialization

Statement

Let A≃A-mod be a finite k-linear abelian category with module model, and let Nr=A∗⊗A− be its Nakayama functor (Left and right Nakayama functors by finite kernel calculus, Nakayama kernels give well-defined adjoint functors). For every finite-dimensional projective left A-module P (Projective modules and the lifting property) and every finite-dimensional left A-module X there is a natural isomorphism DHom⁡A(P,X)≅Hom⁡A(X,A∗⊗AP)=Hom⁡A(X,Nr(P)), where D=Hom⁡k(−,k) is the k-dual (Linear functionals and the algebraic dual V∗=L(V,F), Finite-dimensional vector space, and its dimension dim⁡FV; infinite-dimensional means having no finite basis, Linear map between vector spaces over the same field). If moreover the module model is supplied with an isomorphism A∗≅A of A-bimodules (the symmetric-algebra condition), then Nr≅1 and Nl≅1; this is a conditional specialization, and no claim is made that every finite-dimensional k-algebra is symmetric or self-injective. No commutativity of A and no choice are used.

Facts & Assumptions

Given: A finite-dimensional unital k-algebra A, a finite k-linear abelian category with module model A≃A-mod, a finite-dimensional projective left A-module P (Projective modules and the lifting property), a finite-dimensional left A-module X, and the Nakayama functors Nr=A∗⊗A−, Nl≅Hom⁡A(A∗,−) of the model (Left and right Nakayama functors by finite kernel calculus, Nakayama kernels give well-defined adjoint functors).

[F1]

The algebra A is an (A,A)-bimodule by left and right multiplication, and A∗=Hom⁡k(A,k) is the (A,A)-bimodule with (a⋅λ)(b)=λ(ba) and (λ⋅a)(b)=λ(ab) (Unital left and right modules over a ring; unqualified module means left module, (S,R)-bimodules and commuting left and right scalar actions, Linear functionals and the algebraic dual V∗=L(V,F)).

[F2]

For a left A-module P, the space Hom⁡A(P,A) is a right A-module under (f⋅a)(p)=f(p)a, since (f⋅a)(bp)=bf(p)a. Its k-dual is a left A-module under (aμ)(f)=μ(f⋅a). Left multiplication on the values of f need not preserve A-linearity when A is noncommutative (Unital left and right modules over a ring; unqualified module means left module, Linear map between vector spaces over the same field).

[F3]

A finite-dimensional left A-module has a finite k-basis, and that finite set generates it as an A-module (through k-linear combinations and the unit), so it is finitely generated (Finite-dimensional vector space, and its dimension dim⁡FV; infinite-dimensional means having no finite basis, Generated submodule, cyclic and finitely generated modules, module basis and free module).

[F4]

For a unital ring B and a left B-module P that is finitely generated and projective, the evaluation map Hom⁡B(P,B)⊗BY→Hom⁡B(P,Y), φ⊗y↦(p↦φ(p)y), is an isomorphism for every left B-module Y, natural in Y; with B=A and Y=X it identifies Hom⁡A(P,X) with Hom⁡A(P,A)⊗AX (The dual-basis isomorphism for a finitely generated projective bimodule, Projective modules and the lifting property).

[F5]

For unital rings A,B, a (B,A)-bimodule M, a left A-module X and a left B-module Y, currying Hom⁡B(M⊗AX,Y)→Hom⁡A(X,Hom⁡B(M,Y)), F↦(x↦(m↦F(m⊗x))), is a bijection natural in X and Y; with B=k, Y=k it gives Hom⁡k(M⊗AX,k)≅Hom⁡A(X,Hom⁡k(M,k)) for every right A-module M (Tensor-Hom adjunction for bimodules over arbitrary unital rings).

[F6]

For a finite-dimensional (A,B)-bimodule Z the k-dual Z∗=Hom⁡k(Z,k) is a (B,A)-bimodule under (b⋅λ)(z)=λ(zb) and (λ⋅a)(z)=λ(az); (−)∗ is a contravariant equivalence carrying isomorphisms to isomorphisms, and the evaluation ev⁡Z:Z→Z∗∗ is a natural isomorphism (Finite module duality is exact with commuting bimodule actions, Linear functionals and the algebraic dual V∗=L(V,F), Natural isomorphism).

[F7]

The balanced tensor product is unital and functorial in the module argument: A⊗AX≅X naturally in X, and a homomorphism of right A-modules induces a natural transformation between the functors −⊗A− (The regular module is a tensor unit: R⊗RN≅N and M⊗RR≅M, Module homomorphisms induce tensor-product homomorphisms functorially).

Proof

technique · direct
1.1givenF1F2

Define γ:A∗⊗AP→Hom⁡A(P,A)∗ by γ(λ⊗p)(f)=λ(f(p)) for λ∈A∗, p∈P and f∈Hom⁡A(P,A). It is well defined: γ(λ⋅a⊗p)(f)=(λ⋅a)(f(p))=λ(af(p)) equals γ(λ⊗ap)(f)=λ(f(ap)) by [F1] and [F2], so the A-balanced relation λ⋅a⊗p=λ⊗ap is respected. It is left A-linear with respect to the left action on A∗⊗AP and the left action (a⋅μ)(f)=μ(f⋅a) on Hom⁡A(P,A)∗: γ(a⋅(λ⊗p))(f)=(a⋅λ)(f(p))=λ(f(p)a) and (a⋅γ(λ⊗p))(f)=γ(λ⊗p)(f⋅a)=λ((f⋅a)(p))=λ(f(p)a), using [F1] and [F2].

2.1step 1.1F5F6algebra

The map γ is an isomorphism: its transpose γ∗:Hom⁡A(P,A)∗∗→(A∗⊗AP)∗ is a bijection, because under the double-duality isomorphism ev⁡:Hom⁡A(P,A)→Hom⁡A(P,A)∗∗ of [F6] and the bijection (A∗⊗AP)∗→Hom⁡A(P,A), θ↦(p↦ev⁡A−1(λ↦θ(λ⊗p))), given by currying [F5] with M=A∗ followed by [F6], the composite corresponds to the identity: γ∗(ev⁡(f))(λ⊗p)=ev⁡(f)(γ(λ⊗p))=λ(f(p)), and p↦ev⁡A−1(λ↦λ(f(p)))=f(p), so the composite sends f to f. Since both comparison maps are bijections, γ∗ is bijective; all spaces here are finite-dimensional, so double duality [F6] reflects the isomorphism of γ∗, and γ is bijective and hence an isomorphism of left A-modules.

3.1step 2.1F3F4F5F6

Since a finite-dimensional module is finitely generated by [F3], the evaluation map of [F4] with B=A and Y=X gives a natural isomorphism Hom⁡A(P,X)≅Hom⁡A(P,A)⊗AX; dualizing it by [F6] and currying by [F5] with M=Hom⁡A(P,A) and Y=k gives a natural isomorphism DHom⁡A(P,X)=Hom⁡k(Hom⁡A(P,X),k)≅Hom⁡A(X,Hom⁡A(P,A)∗), and composing with Hom⁡A(X,γ−1) from step 2.1 gives the natural isomorphism DHom⁡A(P,X)≅Hom⁡A(X,A∗⊗AP)=Hom⁡A(X,Nr(P)). All three isomorphisms are natural in X (and in P, since the evaluation formula of [F4] and the formula of γ are natural in P), so the composite is a natural isomorphism.

4.1step 3.1F1F7∎

Assume now that the model carries an isomorphism α:A∗→A of A-bimodules. Then Nr=A∗⊗A−≅A⊗A−≅1, the first natural isomorphism induced by α through functoriality in the first variable and the second the unit isomorphism of [F7]; and Nl≅Hom⁡A(A∗,−)≅Hom⁡A(A,−)≅1, where precomposition with α and with α−1 are mutually inverse natural bijections Hom⁡A(A,−)→Hom⁡A(A∗,−) and f↦f(1A) with inverse x↦(a↦ax) is the natural isomorphism Hom⁡A(A,−)≅1 supplied by the module axioms of [F1]. This is a conditional specialization: only the supplied bimodule isomorphism is used, and no claim is made that an arbitrary finite-dimensional k-algebra is symmetric or self-injective. Only the finite-dimensional data and finitely many operations enter, so no commutativity of A and no choice are used.

Depends on

Used by

Nothing in the library uses this result yet.

Dependency tree · two levels

73 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