Lemma 6.7 — restrictions along , realized in the Gelfand–Tsetlin bases
ProvedGL2QFT.lemma_6_7branching-rulesgl2p2o-gran-per-chapterp2o-plan-paperp2o-strengthenedp2o-v1representation-theory
In the explicit models and Gelfand–Tsetlin bases of §6.1–6.2 the restrictions of Lemma 6.7 hold as equalities of matrices.
- (102) .
- (103) is diagonal in the basis (108), with entries ; thus .
- (104) restricts to on .
- (105) On , is block diagonal: on and on the .
- (106) On , () is block diagonal: on , on , on the .
- (107) On , with .
- (Multiplicity-free) The characters of in (103) are pairwise distinct. The summands and of (105) are inequivalent as representations of , and so are any two of the summands , , of (106), .
These branching rules show that the bases of §6.2 are subgroup-adapted along , which is what the induction method of §3 and the Mackey circuit of §5 require.
Formalization Note "" is realized by the identity map in the Gelfand–Tsetlin bases: the restricted matrix equals the block matrix. Multiplicity-freeness is expressed by the inequivalence of the summands.
Differences from the source:
- generalization: isomorphisms of restricted representations (102)–(107) → equalities of the restricted matrices with the block-diagonal matrices in the Gelfand–Tsetlin bases; reason: an explicit decomposition implies the isomorphism, and the paper's §6.2 asserts exactly these adapted bases
- reading: "restriction along the chain is multiplicity-free" → the -characters in (103) are pairwise distinct and the -summands in (105), (106) are pairwise inequivalent; reason: given the decompositions 1–6, multiplicity-freeness means exactly that the summands are pairwise inequivalent
- encoding: , , → the exponent labels , , ; reason: Remark 6.2, labels are exponents
- reading: (106) for every and (107) for every regular → (106) for the representative with (exponent order) in the basis (110), and (107) for the chosen representative of each orbit ; reason: Lemma 6.7 is about the irreducible representations of Table 2, i.e. isomorphism classes, and p.44 represents each principal series by with ( by Lemma 6.5, and the right side of (106) is symmetric in ); and have the same matrices in this model
Preamble
import Definitions.Def_GL2QFT_Irreps open Classical
Formal statement
namespace GL2QFT
open FieldData
/-- arXiv:2608.28573 (held version, sha256 ecd45073…), **Lemma 6.7**, p.37 (PDF p.38): "The following
restrictions of irreducible representations hold:
`Res^B_T χ_{α,β} = χ_{α,β}` (102), `Res^B_T ρ_γ = ⊕_{δ∈X_q} χ_{δ,δ⁻¹γ}` (103),
`Res^G_B det_α = χ_{α,α}` (104), `Res^G_B St_α = χ_{α,α} ⊕ ρ_{α²}` (105),
`Res^G_B I_{α,β} = χ_{α,β} ⊕ χ_{β,α} ⊕ ρ_{αβ}`, `α ≠ β` (106),
`Res^G_B π_θ = ρ_γ`, `θ|_{F*_q} = χ_γ` (107).
In particular, restriction along the chain `T < B < GL₂(F_q)` is multiplicity-free."
Stated in the explicit models and Gelfand–Tsetlin bases, so that every `≅` is realized by the
basis (as in §6.2): for `g ∈ T` (diagonal) resp. `g ∈ B` (upper triangular), the matrix of the
restriction is block diagonal with the stated blocks, in this order:
1. (102) `χ_{α,β}(diag(a, d)) = χ_α(a)χ_β(d)`;
2. (103) `ρ_γ(diag(a, d))` is diagonal in (108) with entries `χ_δ(a) χ_{δ⁻¹γ}(d)`;
3. (104) `R_{det_α}(g) = χ_{α,α}(g)` on `B`;
4. (105) `R_{St_α}(g)` on `B`: `χ_{α,α}` on `|v^⊥⟩`, `ρ_{α²}` (in (108)) on the `|w_δ⟩`, zero across;
5. (106) `R_{I_{α,β}}(g)` on `B`, `α ≺ β`: `χ_{α,β}` on `|∞⟩`, `χ_{β,α}` on `|v₀⟩`, `ρ_{αβ}` on the
`|w_δ⟩`, zero across;
6. (107) `R_{π_θ}(g) = ρ_γ(g)` on `B`, `γ = γ(k)`;
7. (multiplicity-free) the `T`-characters in (103) are pairwise distinct, and the `B`-summands in
(105) and in (106) are pairwise inequivalent as representations of `B`. -/
theorem lemma_6_7 (D : FieldData) :
(∀ (α β : D.Lab) (g : D.G), (g : Matrix (Fin 2) (Fin 2) D.Fq) 1 0 = 0 →
(g : Matrix (Fin 2) (Fin 2) D.Fq) 0 1 = 0 →
D.RB (.inl (α, β)) (g : Matrix (Fin 2) (Fin 2) D.Fq) () () =
D.chi α ((g : Matrix (Fin 2) (Fin 2) D.Fq) 0 0) * D.chi β ((g : Matrix (Fin 2) (Fin 2) D.Fq) 1 1)) ∧
(∀ (γ : D.Lab) (g : D.G) (δ δ' : D.Lab), (g : Matrix (Fin 2) (Fin 2) D.Fq) 1 0 = 0 →
(g : Matrix (Fin 2) (Fin 2) D.Fq) 0 1 = 0 →
D.RB (.inr γ) (g : Matrix (Fin 2) (Fin 2) D.Fq) δ δ' =
if δ = δ' then D.chi δ ((g : Matrix (Fin 2) (Fin 2) D.Fq) 0 0) *
D.chi (γ - δ) ((g : Matrix (Fin 2) (Fin 2) D.Fq) 1 1) else 0) ∧
(∀ (α : D.Lab) (g : D.G), (g : Matrix (Fin 2) (Fin 2) D.Fq) 1 0 = 0 →
D.R (.inl α) g () () = D.RB (.inl (α, α)) (g : Matrix (Fin 2) (Fin 2) D.Fq) () ()) ∧
(∀ (α : D.Lab) (g : D.G), (g : Matrix (Fin 2) (Fin 2) D.Fq) 1 0 = 0 →
∀ i j : Option D.Lab, D.R (.inr (.inl α)) g i j =
match i, j with
| none, none => D.RB (.inl (α, α)) (g : Matrix (Fin 2) (Fin 2) D.Fq) () ()
| some δ, some δ' => D.RB (.inr (2 * α)) (g : Matrix (Fin 2) (Fin 2) D.Fq) δ δ'
| _, _ => 0) ∧
(∀ (ab : {ab : D.Lab × D.Lab // ab.1.val < ab.2.val}) (g : D.G),
(g : Matrix (Fin 2) (Fin 2) D.Fq) 1 0 = 0 →
∀ i j : Option (Option D.Lab), D.R (.inr (.inr (.inl ab))) g i j =
match i, j with
| none, none => D.RB (.inl (ab.1.1, ab.1.2)) (g : Matrix (Fin 2) (Fin 2) D.Fq) () ()
| some none, some none => D.RB (.inl (ab.1.2, ab.1.1)) (g : Matrix (Fin 2) (Fin 2) D.Fq) () ()
| some (some δ), some (some δ') =>
D.RB (.inr (ab.1.1 + ab.1.2)) (g : Matrix (Fin 2) (Fin 2) D.Fq) δ δ'
| _, _ => 0) ∧
(∀ (k : {k : D.Lab2 // D.IsCuspLabel k}) (g : D.G), (g : Matrix (Fin 2) (Fin 2) D.Fq) 1 0 = 0 →
∀ i j : D.Lab, D.R (.inr (.inr (.inr k))) g i j =
D.RB (.inr (D.restrLab k.1)) (g : Matrix (Fin 2) (Fin 2) D.Fq) i j) ∧
((∀ γ δ δ' : D.Lab, δ ≠ δ' → ∃ g : D.G, (g : Matrix (Fin 2) (Fin 2) D.Fq) 1 0 = 0 ∧
(g : Matrix (Fin 2) (Fin 2) D.Fq) 0 1 = 0 ∧
D.chi δ ((g : Matrix (Fin 2) (Fin 2) D.Fq) 0 0) * D.chi (γ - δ) ((g : Matrix (Fin 2) (Fin 2) D.Fq) 1 1) ≠
D.chi δ' ((g : Matrix (Fin 2) (Fin 2) D.Fq) 0 0) *
D.chi (γ - δ') ((g : Matrix (Fin 2) (Fin 2) D.Fq) 1 1)) ∧
(∀ α : D.Lab, ¬ RepEquivOn {g : D.G | (g : Matrix (Fin 2) (Fin 2) D.Fq) 1 0 = 0}
(fun g : D.G => D.RB (.inl (α, α)) (g : Matrix (Fin 2) (Fin 2) D.Fq))
(fun g : D.G => D.RB (.inr (2 * α)) (g : Matrix (Fin 2) (Fin 2) D.Fq))) ∧
(∀ α β : D.Lab, α ≠ β →
¬ RepEquivOn {g : D.G | (g : Matrix (Fin 2) (Fin 2) D.Fq) 1 0 = 0}
(fun g : D.G => D.RB (.inl (α, β)) (g : Matrix (Fin 2) (Fin 2) D.Fq))
(fun g : D.G => D.RB (.inl (β, α)) (g : Matrix (Fin 2) (Fin 2) D.Fq)) ∧
¬ RepEquivOn {g : D.G | (g : Matrix (Fin 2) (Fin 2) D.Fq) 1 0 = 0}
(fun g : D.G => D.RB (.inl (α, β)) (g : Matrix (Fin 2) (Fin 2) D.Fq))
(fun g : D.G => D.RB (.inr (α + β)) (g : Matrix (Fin 2) (Fin 2) D.Fq)) ∧
¬ RepEquivOn {g : D.G | (g : Matrix (Fin 2) (Fin 2) D.Fq) 1 0 = 0}
(fun g : D.G => D.RB (.inl (β, α)) (g : Matrix (Fin 2) (Fin 2) D.Fq))
(fun g : D.G => D.RB (.inr (α + β)) (g : Matrix (Fin 2) (Fin 2) D.Fq)))) := by sorry
end GL2QFT
Source
Bruinsma, Posta, Stokvis, Grinko, Ozols, Quantum Fourier transform toolbox, arXiv:2608.28573 (held version, sha256 ecd45073…), Lemma 6.7 eqs. (102)–(107), p.37 (PDF p.38); §6.2.1–6.2.2 pp.37–39
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.