Alphabeta Math
TheoremStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedprecheck pass
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.

Murnaghan stability for Kronecker coefficients

Statement

Fix partitions α,β,γ and work over a field k of characteristic 0, with Vλ denoting the irreducible kSn-module of shape λ⊢n (Specht modules classify the irreducibles over every field of characteristic zero). Put m(γ):=max⁡{∣α∣+α1, ∣β∣+β1, ∣γ∣+γ1} and, for n≥m(γ), let gn(α,β,γ):=gα[n],β[n]γ[n] be the multiplicity of Vγ[n] in Vα[n]⊗Vβ[n] (Kronecker coefficients and the internal tensor product, Padded Kronecker coefficient sequences); set gn(α,β,γ):=0 for n<m(γ), where the padded sequences are not partitions.

Then the sequence n↦gn(α,β,γ) is eventually constant. More precisely, there is a finite set S(α,β) of partitions and nonnegative integers gˉα,βγ for γ∈S(α,β), with gˉα,βγ:=0 for γ∉S(α,β), such that gn(α,β,γ)=gˉα,βγ for all sufficiently large n and every partition γ. No explicit bound on the stable rank is asserted here; the sharp global bound is a separate claim.

Facts & Assumptions

Given: Partitions α,β,γ, a field k of characteristic 0, and the FI-module machinery over k.

[F1]

There is a finitely generated FI-module V(α) over k with V(α)n≅Vα[n] for n≥∣α∣+α1 and V(α)n=0 below; it has weight ∣α∣ and stability degree α1 (Each padded irreducible family is realised by a finitely generated FI-module).

[F2]

The pointwise tensor product of finitely generated FI-modules is finitely generated, with (V⊗W)S=VS⊗kWS; if V is generated in degree ≤m1 and W in degree ≤m2, then V⊗W is generated in degree ≤m1+m2 (Tensor products of finitely generated FI-modules are finitely generated).

[F3]

A finitely generated FI-module over a characteristic-zero field with finite-dimensional values is uniformly representation stable, so the multiplicity of each admissible padded irreducible family is constant in a common stable range (Finite generation is equivalent to uniform representation stability, Uniform representation stability for consistent sequences of Sn-modules).

[F4]

For partitions λ≠λ′ of possibly different sizes and n≥max⁡(∣λ∣+λ1,∣λ′∣+λ1′), the irreducible modules Vλ[n] and Vλ′[n] of kSn are nonisomorphic; and Vμ⊗Vν≅⨁γgμ,νγVγ with gμ,νγ the Kronecker coefficient (Kronecker coefficients and the internal tensor product, First-row padding of a partition, Specht modules classify the irreducibles over every field of characteristic zero).

[F5]

Generation in a finite degree d bounds the weight by d: only tails λ with ∣λ∣≤d can occur (Bounded generation degree bounds surjectivity degree and weight, Weight of an FI-module over a field of characteristic zero).

Proof

technique · direct
1.1F1

By [F1] the assignment n↦Vα[n] (and 0 below the threshold) is a finitely generated FI-module V(α) over k with finite-dimensional values, and likewise V(β).

2.1F2F4step 1.1

By [F2] the pointwise tensor product T:=V(α)⊗V(β) is finitely generated and has finite-dimensional values; at every rank where both paddings are admissible its value is Vα[n]⊗kVβ[n]. Its tensor multiplicities are the usual complex Kronecker coefficients: by the rational-model construction in Specht modules classify the irreducibles over every field of characteristic zero, Vμ=k⊗QSQμ. The map (c⊗u)⊗(e⊗v)↦ce⊗(u⊗v) identifies its tensor product with the scalar extension of the rational tensor product. Extend a rational Specht decomposition to k and to C; tensoring carries a finite direct sum to the same sum of extended Specht modules (by the displayed map on bases), so the two fields have identical multiplicities.

3.1F3F5step 2.1

Since T is finitely generated, it is generated in some finite degree d. By [F5] only the finitely many partitions λ with ∣λ∣≤d can label constituents. By [F3] their multiplicities are constant from some rank N onward. Increase N to make all these finitely many labels admissible. Then Tn≅⨁λ∈ScλVλ[n] for n≥N, where S is the finite set of labels with positive multiplicity; put gˉα,βγ:=cγ on S and zero outside S.

4.1F4step 3.1

Put N0:=max⁡{N, ∣α∣+α1, ∣β∣+β1, max⁡({0}∪{∣λ∣+λ1:λ∈S})}, finite because S is finite. For n≥N0 the equality of step 3.1 holds with every Vγ[n], γ∈S, an honest irreducible of kSn; and the rank-n value also decomposes by [F4] as ⨁γgα[n],β[n]γ[n]Vγ[n], the sum over admissible tails γ with n≥∣γ∣+γ1, which label all partitions of n uniquely by deleting their first row. By the nonisomorphism clause of [F4], the two decompositions have the same labels and multiplicities: gn(α,β,γ)=gˉα,βγ for every partition γ and every n≥N0, where for γ∉S this says gn(α,β,γ)=0.

5.1givenstep 4.1

The case γ∉S and n<m(γ) also gives gn(α,β,γ)=0 by the convention in the statement, so making N0 larger if necessary the equality gn(α,β,γ)=gˉα,βγ holds for all n≥N0 and all partitions γ; this is eventual constancy with the finite support S(α,β):=S. The empty partition is included: V∅[n] is the trivial module, the thresholds ∣∅∣+∅1=0 cause no special case, and the argument did not use any bound on N0 sharper than existence.

6.1F1F2F3step 4.1∎

No choice principle is used: the FI-module suppliers are Z-linear constructions on finite-dimensional representation categories, the finite family S and its multiplicities come from [F3], and the comparison of irreducibles in step 4.1 is a finite-dimensional decomposition statement over k; the statement is asserted for every field of characteristic 0.

Depends on

Used by

Dependency tree · two levels

49 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