The ‖·‖_ψ toolkit #
The trivial half: ‖x y Ω‖ ≤ ‖x‖ ‖y Ω‖.
A self-adjoint central element acts on Ω through the commutant: w Ω = (J w J) Ω
for w ∈ M self-adjoint and commuting with R.
HJX Lemma 2.5(ii): ‖y w Ω‖ ≤ ‖w‖ ‖y Ω‖ for w self-adjoint in the centralizer.
Conjugation by a unitary of the centralizer is ‖·‖_ψ-isometric.
Multiplying on the right by a unitary of the centralizer is ‖·‖_ψ-isometric.
‖[w, y]Ω‖ ≤ 2‖w‖‖yΩ‖ for w self-adjoint in the centralizer.
‖[w^{k+1}, y]Ω‖ ≤ (k+1)‖w‖^k‖[w,y]Ω‖ for w self-adjoint in the centralizer.
HJX Lemma 2.6(ii): ‖[e^{itw}, y]Ω‖ ≤ |t| e^{|t|‖w‖}‖[w,y]Ω‖.
Left-bounded elements #
x is left bounded with constant c: ψ(x* y x) ≤ c ψ(y) for 0 ≤ y ∈ M
(HJX Lemma 2.5).
Equations
Instances For
‖y x Ω‖ ≤ √c ‖y Ω‖ for a left-bounded x.
The commutator estimate: ‖[y, x] Ω‖ ≤ (√c + ‖x‖) ‖y Ω‖ for a left-bounded x.
The Gaussian-smeared elements are left bounded (E4.5, HJX Lemma 2.5(i)).
The bounded logarithms b_n = 2^{-n} a_n #
b_n = −i Log λ(2^{-n}), so that a_n = 2^n b_n.
Equations
Instances For
e^{ikb_n} = λ(k2^{-n}) for every integer k.
HJX Lemma 2.2 in this model: Ω̂ is Haar distributed for b_n, because the vectors
λ(k2^{-n})Ω̂ = δ_{k2^{-n}} ⊗ Ω are pairwise orthogonal.
HJX Lemma 2.6(i): ‖[b_n, x]Ω̂‖ → 0 #
‖[λ(q), x]Ω̂‖ = ‖xΩ̂ − Δ̂^{-iq}(xΩ̂)‖, because λ(q) is unitary and
λ(-q)xλ(q) = σ̂_{-q}(x).
HJX Lemma 2.6(i): ‖[b_n, x] Ω̂‖ → 0 for every x ∈ ℛ.
HJX Lemma 2.6(iii) and the density theorem #
HJX Lemma 2.6(iii), pointwise in t: on ℛ,
‖σ^{ξ_n}_t(x)Ω̂ − xΩ̂‖ ≤ ‖Δ̂^{it}(xΩ̂) − xΩ̂‖ + ‖[e^{ita_n}, x]Ω̂‖.
The averaging bound: a uniform bound on [0, 2^{-n}] bounds ‖(Φ_n x − x)Ω̂‖.
Haagerup's density theorem (the final lemma of HJX §2): Φ_n(x)Ω̂ → xΩ̂ for every
x ∈ ℛ. Since Φ_n(x) ∈ ℛ_n (E6.4), the union ⋃_n ℛ_n is ‖·‖_{ψ̂}-dense in ℛ.
⋃_n ℛ_n is ‖·‖_{ψ̂}-dense in ℛ: every x ∈ ℛ is ‖·‖_{ψ̂}-approximated by
Φ_n(x) ∈ ℛ_n.