Alphabeta Math
TheoremStatement: 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.

Peter-Weyl decomposition of the regular representation

Statement

Assume the Axiom of Choice (The Axiom of Choice). Let K be a compact Hausdorff group with normalized Haar probability μ and let λ,ρ be the left and right regular unitary representations of K on L2(K,μ;C) (Left and right regular unitary representations of an LCH group; K is unimodular so ΔK≡1). For π∈K^ let Mπ⊆C(K) be the span of all matrix coefficients of π (Matrix coefficient of a unitary representation). For the fixed representative π of its class, with fixed orthonormal basis e1π,…,edππ, let Jπ:Hπ→Hπ be the coordinate conjugation Jπ(∑iaieiπ)=∑iai‾eiπ, and define the conjugate representation π‾(k):=Jππ(k)Jπ. Then:

  1. Mπ is finite dimensional of dimension dπ2 and is the closed span of the block {uijπ}i,j of The normalized irreducible matrix coefficient family; distinct Mπ,Mσ are orthogonal, and L2(K)=⨁^π∈K^Mπ (Hilbert direct sums of unitary representations).
  2. The two-sided action on coefficients is ρ(g)cv,wπ=cπ(g)v,wπ and λ(g)cv,wπ=cv,π(g)wπ for all g∈K and v,w. Consequently ρ∣Mπ≅dπ π and λ∣Mπ≅dπ π‾ as unitary representations.
  3. Hence ρ≅⨁^π∈K^dπ π and λ≅⨁^π∈K^dπ π‾; the assignment π↦π‾ induces a bijection of K^ with dπ‾=dπ, so also λ≅⨁^π∈K^dπ π: the left regular representation is the Hilbert direct sum of dπ copies of π, and on each coefficient block the right action is on the input-vector factor Hπ, while the left action is on its conjugate factor.

Facts & Assumptions

[F1]

The normalized family B=(uijπ) is an orthonormal basis of L2(K), with uijπ=dπ ceiπ,ejππ and cv,wπ(k)=⟨π(k)v,w⟩; the conjugates of the coefficients of a single class satisfy the orthogonality relations of the next fact. (The normalized matrix coefficients form an orthonormal basis of L2(K), The normalized irreducible matrix coefficient family, Matrix coefficient of a unitary representation)

[F2]

Schur orthogonality for compact groups: cv,wπ and cv′,w′σ are orthogonal in L2(K) when π,σ are inequivalent irreducibles, and ∫Kcv,wπcv′,w′π‾ dμ=dπ−1⟨v,v′⟩⟨w,w′⟩‾ for one class. (Schur orthogonality for general compact groups)

[F3]

The left and right regular representations are λ(g)h(x)=h(g−1x) and ρ(g)h(x)=h(xg) on L2(K) (unimodularity removes the modular factor), and both are strongly continuous unitary representations. (Left and right regular unitary representations of an LCH group, The regular representations are unitary, strongly continuous, and the left one is faithful, Compact, discrete and abelian groups are unimodular)

[F4]

If a Hilbert space is the orthogonal Hilbert direct sum of closed invariant subspaces on which the restrictions are unitarily equivalent to given representations, then the whole representation is the Hilbert direct sum of those subrepresentations; equivalently, ρ∣Mπ≅dππ on each block gives ρ≅⨁^πdππ. (Hilbert direct sums of unitary representations)

[F5]

Dimension is additive over direct sums and equal for a vector space and its image under a linear isomorphism. (Finite-dimensional vector space, and its dimension dim⁡FV; infinite-dimensional means having no finite basis, Hilbert direct sums of unitary representations)

Proof

Given: AC, a compact Hausdorff group K with normalized Haar probability μ, its unitary dual K^ with fixed representatives and orthonormal bases, and the two regular representations λ,ρ.

1.1F1F2F5

For a class π of dimension dπ, the dπ2 functions cei,ejπ are linearly independent: if ∑i,jλijcei,ejπ=0, then by [F2] its squared L2 norm is ∑i,j∣λij∣2/dπ, which must vanish, so every λij=0; hence Mπ=span⁡{cei,ejπ:i,j} has dimension dπ2, and because uijπ=dπcei,ejπ it is also the span of the block {uijπ}i,j; for inequivalent classes the coefficients are pairwise orthogonal by [F2], so Mπ⊥Mσ. The closed span of ⋃πMπ contains every uijπ and hence the closed span of the Hilbert basis B, which is L2(K) by [F1]; therefore L2(K)=⨁^π∈K^Mπ is an orthogonal Hilbert-space direct sum of these blocks.

2.1F1F2F3F4step 1.1

For all g∈K and v,w∈Hπ, [F3] gives ρ(g)cv,wπ(x)=cv,wπ(xg)=⟨π(x)π(g)v,w⟩=cπ(g)v,wπ(x) and λ(g)cv,wπ(x)=cv,wπ(g−1x)=⟨π(g−1x)v,w⟩=⟨π(x)v,π(g)w⟩=cv,π(g)wπ(x), which are the stated action formulas. Fix the orthonormal basis e1,…,ed of Hπ and put Hj:=span⁡i{cei,ejπ}; each Hj has dimension d by the linear independence in step 1.1, satisfies ρ(g)Hj⊆Hj by the first formula, and is the image of Hπ under the linear map Vj(v):=cv,ejπ with ρ(g)Vj=Vjπ(g); since ∥cv,ejπ∥22=d−1∥v∥2 by [F2], d Vj is a unitary intertwiner Hπ→Hj. The sum ∑jHj equals Mπ and ∑jdim⁡Hj=d2=dim⁡Mπ, and [F2] makes distinct Hj orthogonal, so Mπ=⨁j=1dHj is an orthogonal direct sum, and ρ∣Mπ≅dπ π by [F4].

3.1F1F2F4step 1.1step 2.1∎

On the same basis define the coordinate conjugation Jπ(∑iaiei)=∑iai‾ei and π‾(g):=Jππ(g)Jπ; then π‾ is a group homomorphism because Jπ2=id⁡, it is unitary because Jπ is a conjugate-linear isometry and π(g) is unitary, it is strongly continuous because Jπ is isometric, and it satisfies dπ‾=dπ and π‾‾=π; moreover π‾ is irreducible exactly when π is, since M↦JπM is a bijection between the closed invariant subspaces of π‾ and those of π. The resulting class is independent of the choices: if T:Hπ→Hπ′ is a unitary intertwiner and J,J′ are the two coordinate conjugations, then J′TJ is a linear unitary intertwiner from JπJ to J′π′J′. Taking T=I also covers a change of basis for the same representation. Conjugating twice returns the original class, so this defines a dimension-preserving involution of K^. For the same basis define Gi:=span⁡j{cei,ejπ}; the second action formula of step 2.1 shows λ(g)cei,ejπ=cei,π(g)ejπ=∑lπlj(g)‾ cei,elπ with πlj(g)=⟨π(g)ej,el⟩, and πlj(g)‾=⟨Jππ(g)ej,el⟩=⟨π‾(g)ej,el⟩ is exactly the (l,j)-entry of π‾(g); the linear map Wi(ej):=cei,ejπ therefore satisfies λ(g)Wi=Wiπ‾(g) and d Wi is a unitary intertwiner Hπ→Gi by [F2], while Mπ=⨁iGi and dim⁡Gi=d as in step 2.1; hence λ∣Mπ≅dπ π‾. Applying [F4] to the block decomposition of step 1.1 gives ρ≅⨁^πdππ and λ≅⨁^πdππ‾, and reindexing the latter sum by the involution π↦π‾ of K^ gives λ≅⨁^πdππ, which completes the proof. The Axiom of Choice is inherited through the fixed representatives and bases and the cited suppliers.

Depends on

Used by

Dependency tree · two levels

78 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