Alphabeta Math
TheoremStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedprecheck passjudge pass (gpt-6-sol)audited 2026-09-30
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.

Shift-orbit bases for graded simple and projective classes

Statement

Let k be a field and A a finite-dimensional unital associative Z-graded k-algebra. A graded-simple module here means a nonzero finite-dimensional graded left A-module whose only graded submodules are 0 and itself. Internal shift acts on graded-simple isomorphism classes by [S]↦[S{r}] for r∈Z. There are finitely many shift orbits of graded-simple isomorphism classes. For each orbit choose a representative Si and a finite graded projective cover pi:Pi↠Si, whose kernel is superfluous among graded submodules. Then

{[Si]}i is a Z[v,v−1]-basis of G0gr(A),{[Pi]}i is a Z[v,v−1]-basis of K0gr(A).

In particular, both modules have the same finite rank, the number of graded simple shift orbits. If A=0, both bases are empty and both groups are zero. No positivity assumption on the grading of A is made.

Facts & Assumptions

Given: The field k, the finite-dimensional unital associative graded algebra A, and finite-dimensional graded left A-modules. Morphisms preserve degree. No axiom of choice is assumed or used.

[F1]

G0gr(A) is the short-exact-sequence group of the category of finite-dimensional graded left A-modules (Graded Grothendieck groups, shift action, and Cartan map).

[F2]

K0gr(A) is the split Grothendieck group of finite graded projectives (Graded Grothendieck groups, shift action, and Cartan map).

[F3]

The shift action is vr[M]=[M{r}] and vr[P]=[P{r}] (Graded Grothendieck groups, shift action, and Cartan map).

[F4]

In an essentially small abelian category of finite-length objects, the simple-object classes form a free abelian basis of G0 (Simple classes freely generate the Grothendieck group of a length category).

[F5]

Every finite-dimensional graded module has a finite decomposition into graded-indecomposable summands, unique up to permutation and degree-zero graded isomorphism (Graded Krull–Schmidt for finite-dimensional graded modules).

[F6]

Every finite-dimensional graded module has a finite graded projective cover with superfluous kernel, and two covers of the same object are isomorphic over it (Finite-dimensional graded algebras have graded projective covers).

[F7]

A degree-zero endomorphism of a finite-dimensional graded-indecomposable module is invertible or nilpotent (Graded Fitting decomposition for degree-zero endomorphisms).

[F8]

A graded module is finite graded projective exactly when it is a degree-zero summand of a finite direct sum of shifts of A; finite sums of such shifts are projective (Finite graded projectives are finite shifted-free summands).

[F9]

A graded projective object lifts degree-zero maps through degree-zero epimorphisms (Finite graded projective modules).

[F10]

In GrMod⁡0(A), kernels, images, cokernels, finite biproducts and exactness are computed degreewise (Graded modules with degree-zero maps form an abelian category).

[F11]

A simple object is nonzero and has no proper nonzero subobject (Simple object).

[F12]

An object has finite length when it admits a composition series (Object of finite length).

[F13]

A composition series is a finite strict chain whose successive quotients are simple (Composition series and composition factors of an object).

[F14]

The split Grothendieck group imposes exactly the relations [P⊕Q]=[P]+[Q] (Split Grothendieck group of an additive category).

[F15]

The free abelian group on a set has its usual universal property (Free abelian group on a set).

[F16]

Internal shift has components M{r}d=Md−r and is invertible, with inverse shift {−r} (Associative graded algebras, bimodules, and internal shifts).

Proof

technique · direct
1.1F10F11F12F13giveninductionchoose

Every finite-dimensional graded module M has finite length. If M=0, the empty chain is a composition series. If M≠0, the dimensions of its proper graded submodules form a nonempty subset of {0,1,…,dim⁡kM−1}; choose a proper graded submodule N of maximal dimension. The quotient M/N is nonzero. Any proper nonzero graded submodule of M/N would lift to a proper graded submodule strictly containing N, contrary to maximality, so M/N is simple. Since dim⁡kN<dim⁡kM, induction gives a composition series of N; appending M/N gives one for M. This uses only a maximum in a finite set of dimensions and one submodule at a time.

1.2F5F8givencases

By graded Krull–Schmidt [F5], write the regular graded module as a finite direct sum A≅⨁j=1mQj of nonzero graded-indecomposable modules; if A=0, take m=0. Each Qj is a direct summand of the shifted free module A{0}, so [F8] makes it a finite graded projective. If A=0, every unital left A-module is zero, so both groups are zero and the empty bases prove the theorem. Henceforth assume A≠0.

1.3F7F9F10F11givenalgebra

Let Q be a nonzero finite-dimensional graded-indecomposable projective and q:Q↠S a degree-zero epimorphism to a graded-simple module. Put K=ker⁡q. If N≤Q is graded and K+N=Q, then q∣N:N↠S is epic. Projectivity [F9] lifts q through q∣N to a degree-zero map g:Q→N. After inclusion N↪Q, let f be the resulting endomorphism. Then qf=q, and induction gives qfn=q for every n≥1, so f is not nilpotent. By graded Fitting [F7], f is invertible. Since im⁡f⊆N, this forces N=Q. Thus K is superfluous among graded submodules and q is a finite graded projective cover.

1.4F6F10F11algebra

Every finite graded projective cover p:P↠S of a graded-simple module is indecomposable. Indeed, if P=U⊕V with both summands nonzero, at least one restriction of p is nonzero and hence surjective; its summand U then satisfies U+ker⁡p=P, contradicting superfluity of ker⁡p. Moreover, S is the unique graded-simple quotient of P up to isomorphism. If q:P↠T is another such quotient and L=ker⁡q, a nonzero graded image q(ker⁡p) must be all of T by simplicity. That would give L+ker⁡p=P, contradicting superfluity. Hence q(ker⁡p)=0, so q factors through P/ker⁡p≅S; the induced nonzero map S→T is an isomorphism.

1.5F16givenchoosealgebra

Every finite-dimensional graded module has finite support: if its dimension is n and it had n+1 distinct nonzero homogeneous components, one nonzero vector from each would be linearly independent. If M≠0 and M≅M{r} by a degree-zero isomorphism, then supp⁡(M)=supp⁡(M)+r. Taking the maximum of this finite nonempty set gives max⁡supp⁡(M)=max⁡supp⁡(M)+r, hence r=0. Thus the shift action is free on the isomorphism classes of nonzero graded simples and nonzero indecomposable projectives; no lower or upper bound on the grading of A is used.

1.6F10givenconstructalgebra

The category of finite-dimensional graded modules is abelian: [F10] makes kernels, cokernels and finite biproducts degreewise, so these objects remain finite-dimensional and the full subcategory inherits the abelian structure. It is essentially small as well. For each finite-support dimension vector on Z, fix the standard graded k-space with those component dimensions; the possible A-actions on it form a set of families of linear maps satisfying the module identities. Every finite-dimensional graded module is isomorphic to one of these models by choosing bases for its finitely many nonzero homogeneous components. The family of all such standard models is a set, and this object-by-object argument makes no simultaneous choice across an arbitrary family.

2.1F1F4step 1.1step 1.6given

By step 1.6, the category of finite-dimensional graded modules is the essentially small abelian category used to define G0gr(A) in [F1]. By step 1.1 all its objects have finite length, so [F4] says that its graded-simple isomorphism classes form a Z-basis of G0gr(A).

2.2F10F11F16step 1.2givenchoosealgebra

If S is graded-simple, take a nonzero homogeneous s∈Sr. The graded submodule As is nonzero, hence is S. The map A{r}→S, a↦as, is degree-zero because 1A has degree r in A{r} and the action preserves degree; it is surjective. Decomposing A{r}≅⨁j=1mQj{r}, at least one restriction to a summand is nonzero and therefore surjective onto the simple module S.

3.1F8F10F16step 1.1step 1.2step 1.3step 1.4step 2.2choose

Each Qj has a graded-simple quotient: a composition series from step 1.1 has a simple final factor. By step 1.3 this quotient map is a projective cover, and by step 1.4 its simple quotient is unique up to isomorphism; denote that isomorphism class by Sj. For any graded-simple S, step 2.2 gives a surjection from some Qj{r} to S. Since shift is invertible [F16], both Qj{r} and Sj{r} retain indecomposability and simplicity, respectively; Qj{r} is finite graded projective by [F8]. The shift of the cover Qj↠Sj is a cover of Sj{r}: [F10] preserves the epimorphism, and shifting back by {−r} preserves the superfluity condition. By steps 1.3–1.4, the quotient S is isomorphic to Sj{r}. Thus the finite list S1,…,Sm meets every graded-simple shift orbit.

4.1F6F8F10F16step 1.3step 1.4step 3.1

Every nonzero finite-dimensional graded-indecomposable projective Q has a graded-simple quotient by step 1.1; step 1.3 makes that quotient map a projective cover. Existence and uniqueness of covers [F6] therefore identify Q with the cover P(S) of its simple quotient. Conversely, step 1.4 shows each P(S) is indecomposable. Shifting a cover gives a cover of the shifted simple, as in step 3.1, so P(S{r})≅P(S){r} by [F6]. If P(S){r}≅P(T), that projective has simple quotients S{r} and T; uniqueness from step 1.4 gives S{r}≅T. Hence projective indecomposable shift orbits are in bijection with graded-simple shift orbits, and there are finitely many.

4.2F1F3F4step 2.1step 3.1step 1.5constructchoose

By steps 2.1 and 3.1, the Z-basis of G0gr(A) is partitioned into finitely many free shift orbits. Choose one simple Si from each orbit. By [F3], vr[Si]=[Si{r}], so the Laurent monomials vr map bijectively to the distinct Z-basis classes in that orbit. The orbit spans therefore form a direct sum of copies of Z[v,v−1], with basis [Si]. This proves the stated finite Laurent basis for G0gr(A).

5.1F1F2F3F5F6F8F14F15step 1.5step 1.6step 4.1construct∎

By step 1.6, the set I of isomorphism classes of nonzero finite-dimensional graded-indecomposable projectives is a set. By graded Krull–Schmidt [F5], each finite graded projective P has a unique finite decomposition into nonzero graded-indecomposable summands Qj. By [F8], P is a degree-zero summand of a finite sum of shifts of A; each Qj, being a summand of P, is also a summand of that finite sum by transitivity of direct summands. Thus [F8] makes every Qj a finite graded projective. Sending P to its multiplicity vector in I is additive under direct sum. By the split-group presentation [F14], it descends to a homomorphism K0gr(A)→Z[I]. The map from the free abelian group [F15] sending each basis vector to its projective class is inverse: one composite fixes each indecomposable basis vector, while the other sends [P] to the sum of its indecomposable classes, which equals [P] by the split relation. Thus I is a Z-basis of K0gr(A). By steps 4.1 and 1.5, this basis is partitioned into finitely many free shift orbits represented by the covers Pi of the chosen Si. Using [F3] as in step 4.2 shows that [Pi] is a finite Z[v,v−1]-basis of K0gr(A). The cover classes are unique up to isomorphism by [F6], so the result is independent of the chosen covers.

Remark

Kleshchev, §2.2, PDF p. 6 (printed p. 7), uses the same positive-shift and homogeneous-map convention. His §2.1 assumes an algebraically closed field; that stronger hypothesis and his ungraded-to-graded simple classification are not used here. The orbit and projective-cover arguments above are proved locally under the stated field hypothesis.

Depends on

Used by

Dependency tree · two levels

46 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