Each result cited below is universally quantified over the data in its own statement.
Letters. We use the following consequences of Words in Unitary Letters: Generators, Signs, Inverse Letters, Lengths and Adjoint Words §letters . For j ∈ [ d ] j\in[d] j ∈ [ d ] the letter j j j has g ( j ) = j g(j)=j g ( j ) = j , ε ( j ) = 1 \varepsilon(j)=1 ε ( j ) = 1 and j − 1 = d + j j^{-1}=d+j j − 1 = d + j , and the letter d + j d+j d + j (which lies in [ 2 d ] [2d] [ 2 d ] and exceeds d d d , by Words in Unitary Letters: Generators, Signs, Inverse Letters, Lengths and Adjoint Words §letters ) has g ( d + j ) = j g(d+j)=j g ( d + j ) = j , ε ( d + j ) = − 1 \varepsilon(d+j)=-1 ε ( d + j ) = − 1 and ( d + j ) − 1 = j (d+j)^{-1}=j ( d + j ) − 1 = j , because j j j is the only natural number m m m with d + j = d + m d+j=d+m d + j = d + m . Consequently, for every letter l ∈ [ 2 d ] l\in[2d] l ∈ [ 2 d ] : g ( l − 1 ) = g ( l ) g(l^{-1})=g(l) g ( l − 1 ) = g ( l ) , ε ( l − 1 ) = − ε ( l ) \varepsilon(l^{-1})=-\varepsilon(l) ε ( l − 1 ) = − ε ( l ) and ( l − 1 ) − 1 = l (l^{-1})^{-1}=l ( l − 1 ) − 1 = l ; and l = g ( l ) l=g(l) l = g ( l ) if ε ( l ) = 1 \varepsilon(l)=1 ε ( l ) = 1 , l = d + g ( l ) l=d+g(l) l = d + g ( l ) if ε ( l ) = − 1 \varepsilon(l)=-1 ε ( l ) = − 1 ; and these two cases are exhaustive, since by Words in Unitary Letters: Generators, Signs, Inverse Letters, Lengths and Adjoint Words §letters the sign ε ( l ) \varepsilon(l) ε ( l ) takes only the values 1 1 1 and − 1 -1 − 1 . For j ∈ [ d ] j\in[d] j ∈ [ d ] and s ∈ { 1 , − 1 } s\in\{1,-1\} s ∈ { 1 , − 1 } we write ℓ ( j , s ) \ell(j,s) ℓ ( j , s ) for the letter j j j if s = 1 s=1 s = 1 and d + j d+j d + j if s = − 1 s=-1 s = − 1 ; thus g ( ℓ ( j , s ) ) = j g(\ell(j,s))=j g ( ℓ ( j , s )) = j , ε ( ℓ ( j , s ) ) = s \varepsilon(\ell(j,s))=s ε ( ℓ ( j , s )) = s , and ℓ ( g ( l ) , ε ( l ) ) = l \ell(g(l),\varepsilon(l))=l ℓ ( g ( l ) , ε ( l )) = l for every letter l l l . For a word w w w of length k k k and a letter l l l , the word w l w\,l w l has length k + 1 k+1 k + 1 and adjoint ( w l ) ∗ = l − 1 w ∗ (w\,l)^{*}=l^{-1}\,w^{*} ( w l ) ∗ = l − 1 w ∗ : by Words in Unitary Letters: Generators, Signs, Inverse Letters, Lengths and Adjoint Words §adjoint and reversal , ( ( w l ) ∗ ) 1 = ( ( w l ) k + 1 ) − 1 = l − 1 ((wl)^{*})_{1}=((wl)_{k+1})^{-1}=l^{-1} (( wl ) ∗ ) 1 = (( wl ) k + 1 ) − 1 = l − 1 and ( ( w l ) ∗ ) i = ( w k + 2 − i ) − 1 = ( w ∗ ) i − 1 ((wl)^{*})_{i}=(w_{k+2-i})^{-1}=(w^{*})_{i-1} (( wl ) ∗ ) i = ( w k + 2 − i ) − 1 = ( w ∗ ) i − 1 for 2 ≤ i ≤ k + 1 2\le i\le k+1 2 ≤ i ≤ k + 1 ; also l ∗ = l − 1 l^{*}=l^{-1} l ∗ = l − 1 and ∅ ∗ = ∅ \varnothing^{*}=\varnothing ∅ ∗ = ∅ . Every nonempty word is of the form w l w\,l w l or is a letter, by Basic Properties of Words: Associativity, Reversal, Finitely Many Factorisations, and Countability §last-letter .
Fact U (unitaries). Let H H H be a complex Hilbert space and T : H → H T:H\to H T : H → H linear. Then T T T is unitary if and only if T ∈ L ( H ) T\in\mathcal{L}(H) T ∈ L ( H ) and T ∗ T = T T ∗ = I T^{*}T=TT^{*}=I T ∗ T = T T ∗ = I . Indeed, let T T T be unitary. By claim 3 of Properties of Unitary Operators and Bounded Linear Maps between Complex Inner Product Spaces: the Least Bound, Operations, the Underlying Real Structure, Adjoints, Completeness and the Quadratic-Form Bound §least-bound , T ∈ L ( H ) T\in\mathcal{L}(H) T ∈ L ( H ) , so T T T has an adjoint T ∗ ∈ L ( H ) T^{*}\in\mathcal{L}(H) T ∗ ∈ L ( H ) by Bounded Linear Maps between Complex Inner Product Spaces: the Least Bound, Operations, the Underlying Real Structure, Adjoints, Completeness and the Quadratic-Form Bound §adjoint , and T T T is the adjoint of T ∗ T^{*} T ∗ by Bounded Linear Maps between Complex Inner Product Spaces: the Least Bound, Operations, the Underlying Real Structure, Adjoints, Completeness and the Quadratic-Form Bound §adjoint-calculus ; by Adjoint of a Linear Map between Complex Inner Product Spaces §adjoint this gives ⟨ u , T ∗ T v ⟩ = ⟨ T u , T v ⟩ = ⟨ u , v ⟩ \langle u,T^{*}Tv\rangle=\langle Tu,Tv\rangle=\langle u,v\rangle ⟨ u , T ∗ T v ⟩ = ⟨ T u , T v ⟩ = ⟨ u , v ⟩ for all u , v ∈ H u,v\in H u , v ∈ H , the second equality being the defining property of T T T . Taking u = T ∗ T v − v u=T^{*}Tv-v u = T ∗ T v − v gives ∥ T ∗ T v − v ∥ 2 = 0 \lVert T^{*}Tv-v\rVert^{2}=0 ∥ T ∗ T v − v ∥ 2 = 0 , so T ∗ T v = v T^{*}Tv=v T ∗ T v = v . Given v ∈ H v\in H v ∈ H , write v = T u v=Tu v = T u by surjectivity; then T T ∗ v = T ( T ∗ T u ) = T u = v TT^{*}v=T(T^{*}Tu)=Tu=v T T ∗ v = T ( T ∗ T u ) = T u = v . Conversely, if T ∈ L ( H ) T\in\mathcal{L}(H) T ∈ L ( H ) and T ∗ T = T T ∗ = I T^{*}T=TT^{*}=I T ∗ T = T T ∗ = I , then every v v v equals T ( T ∗ v ) T(T^{*}v) T ( T ∗ v ) , and ⟨ T u , T v ⟩ = ⟨ u , T ∗ T v ⟩ = ⟨ u , v ⟩ \langle Tu,Tv\rangle=\langle u,T^{*}Tv\rangle=\langle u,v\rangle ⟨ T u , T v ⟩ = ⟨ u , T ∗ T v ⟩ = ⟨ u , v ⟩ as above; so T T T is unitary.
Proof of clause 1 (law). Fix ( H , M , Ω ) (H,M,\Omega) ( H , M , Ω ) and V V V as in clause 1, and write τ = τ M \tau=\tau_{M} τ = τ M ; by Tracial W*-Probability Spaces §trace and Cyclic Tracial Operator Algebras and Their Traces §trace , τ ( T ) = ⟨ Ω , T Ω ⟩ \tau(T)=\langle\Omega,T\Omega\rangle τ ( T ) = ⟨ Ω , T Ω ⟩ for T ∈ M T\in M T ∈ M . Each V j V_{j} V j is unitary, so V j ∈ L ( H ) V_{j}\in\mathcal{L}(H) V j ∈ L ( H ) by claim 3 of Properties of Unitary Operators and Bounded Linear Maps between Complex Inner Product Spaces: the Least Bound, Operations, the Underlying Real Structure, Adjoints, Completeness and the Quadratic-Form Bound §least-bound , and hence V j V_{j} V j has an adjoint V j ∗ ∈ L ( H ) V_{j}^{*}\in\mathcal{L}(H) V j ∗ ∈ L ( H ) by Bounded Linear Maps between Complex Inner Product Spaces: the Least Bound, Operations, the Underlying Real Structure, Adjoints, Completeness and the Quadratic-Form Bound §adjoint , as in Fact U.
(i) V ( u v ) = V ( u ) V ( v ) V^{(uv)}=V^{(u)}V^{(v)} V ( uv ) = V ( u ) V ( v ) for all u , v ∈ W 2 d u,v\in W_{2d} u , v ∈ W 2 d . If u = ∅ u=\varnothing u = ∅ or v = ∅ v=\varnothing v = ∅ this holds because V ( ∅ ) = I V^{(\varnothing)}=I V ( ∅ ) = I and ∅ v = v \varnothing v=v ∅ v = v , u ∅ = u u\varnothing=u u ∅ = u (Words over a Finite Alphabet: the Empty Word, Concatenation and Reversal §concatenation ); otherwise the letters of u v uv uv are those of u u u followed by those of v v v (Words over a Finite Alphabet: the Empty Word, Concatenation and Reversal §concatenation ), and composition of maps is associative.
(ii) ( V ( l ) ) ∗ = V ( l − 1 ) (V^{(l)})^{*}=V^{(l^{-1})} ( V ( l ) ) ∗ = V ( l − 1 ) for every letter l l l . If ε ( l ) = 1 \varepsilon(l)=1 ε ( l ) = 1 , then V ( l ) = V g ( l ) V^{(l)}=V_{g(l)} V ( l ) = V g ( l ) and, as ε ( l − 1 ) = − 1 \varepsilon(l^{-1})=-1 ε ( l − 1 ) = − 1 and g ( l − 1 ) = g ( l ) g(l^{-1})=g(l) g ( l − 1 ) = g ( l ) , V ( l − 1 ) = V g ( l ) ∗ V^{(l^{-1})}=V_{g(l)}^{*} V ( l − 1 ) = V g ( l ) ∗ . If ε ( l ) = − 1 \varepsilon(l)=-1 ε ( l ) = − 1 , then V ( l ) = V g ( l ) ∗ V^{(l)}=V_{g(l)}^{*} V ( l ) = V g ( l ) ∗ , whose adjoint is V g ( l ) = V ( l − 1 ) V_{g(l)}=V^{(l^{-1})} V g ( l ) = V ( l − 1 ) by Bounded Linear Maps between Complex Inner Product Spaces: the Least Bound, Operations, the Underlying Real Structure, Adjoints, Completeness and the Quadratic-Form Bound §adjoint-calculus .
(iii) ( V ( w ) ) ∗ = V ( w ∗ ) (V^{(w)})^{*}=V^{(w^{*})} ( V ( w ) ) ∗ = V ( w ∗ ) for every w ∈ W 2 d w\in W_{2d} w ∈ W 2 d . For w = ∅ w=\varnothing w = ∅ both sides are I I I (Bounded Linear Maps between Complex Inner Product Spaces: the Least Bound, Operations, the Underlying Real Structure, Adjoints, Completeness and the Quadratic-Form Bound §adjoint-calculus ); for a letter this is (ii); and if it holds for w w w of length k k k , then for a letter l l l , by (i), Bounded Linear Maps between Complex Inner Product Spaces: the Least Bound, Operations, the Underlying Real Structure, Adjoints, Completeness and the Quadratic-Form Bound §adjoint-calculus and (ii), ( V ( w l ) ) ∗ = ( V ( w ) V ( l ) ) ∗ = V ( l − 1 ) V ( w ∗ ) = V ( l − 1 w ∗ ) = V ( ( w l ) ∗ ) (V^{(wl)})^{*}=(V^{(w)}V^{(l)})^{*}=V^{(l^{-1})}V^{(w^{*})}=V^{(l^{-1}w^{*})}=V^{((wl)^{*})} ( V ( wl ) ) ∗ = ( V ( w ) V ( l ) ) ∗ = V ( l − 1 ) V ( w ∗ ) = V ( l − 1 w ∗ ) = V (( wl ) ∗ ) . Induction on the length gives (iii).
(iv) V ( l ) V ( l − 1 ) = I V^{(l)}V^{(l^{-1})}=I V ( l ) V ( l − 1 ) = I for every letter l l l : by Fact U, V g ( l ) V g ( l ) ∗ = I V_{g(l)}V_{g(l)}^{*}=I V g ( l ) V g ( l ) ∗ = I and V g ( l ) ∗ V g ( l ) = I V_{g(l)}^{*}V_{g(l)}=I V g ( l ) ∗ V g ( l ) = I , and by (ii) the product V ( l ) V ( l − 1 ) V^{(l)}V^{(l^{-1})} V ( l ) V ( l − 1 ) is the first of these if ε ( l ) = 1 \varepsilon(l)=1 ε ( l ) = 1 and the second if ε ( l ) = − 1 \varepsilon(l)=-1 ε ( l ) = − 1 .
Now we check the four properties of Laws of d-Tuples of Unitaries §law . Normalised: λ V ( ∅ ) = τ ( I ) = ⟨ Ω , Ω ⟩ = ∥ Ω ∥ 2 = 1 \lambda_{V}(\varnothing)=\tau(I)=\langle\Omega,\Omega\rangle=\lVert\Omega\rVert^{2}=1 λ V ( ∅ ) = τ ( I ) = ⟨ Ω , Ω ⟩ = ∥ Ω ∥ 2 = 1 by Cyclic Tracial Operator Algebras and Their Traces §cyclic . Cancellation: by (i) and (iv), V ( u l l − 1 v ) = V ( u ) V ( l ) V ( l − 1 ) V ( v ) = V ( u ) V ( v ) = V ( u v ) V^{(u\,l\,l^{-1}\,v)}=V^{(u)}V^{(l)}V^{(l^{-1})}V^{(v)}=V^{(u)}V^{(v)}=V^{(uv)} V ( u l l − 1 v ) = V ( u ) V ( l ) V ( l − 1 ) V ( v ) = V ( u ) V ( v ) = V ( uv ) , so λ V ( u l l − 1 v ) = λ V ( u v ) \lambda_{V}(u\,l\,l^{-1}\,v)=\lambda_{V}(uv) λ V ( u l l − 1 v ) = λ V ( uv ) . Cyclic: by (i) and Cyclic Tracial Operator Algebras and Their Traces §tracial , λ V ( u v ) = τ ( V ( u ) V ( v ) ) = τ ( V ( v ) V ( u ) ) = λ V ( v u ) \lambda_{V}(uv)=\tau(V^{(u)}V^{(v)})=\tau(V^{(v)}V^{(u)})=\lambda_{V}(vu) λ V ( uv ) = τ ( V ( u ) V ( v ) ) = τ ( V ( v ) V ( u ) ) = λ V ( vu ) . Positive: let F ⊆ W 2 d F\subseteq W_{2d} F ⊆ W 2 d be nonempty and finite, and put ξ v = V ( v ) Ω \xi_{v}=V^{(v)}\Omega ξ v = V ( v ) Ω for v ∈ F v\in F v ∈ F . By (i), (iii), and since V ( v ) V^{(v)} V ( v ) is the adjoint of ( V ( v ) ) ∗ (V^{(v)})^{*} ( V ( v ) ) ∗ (Bounded Linear Maps between Complex Inner Product Spaces: the Least Bound, Operations, the Underlying Real Structure, Adjoints, Completeness and the Quadratic-Form Bound §adjoint-calculus , Adjoint of a Linear Map between Complex Inner Product Spaces §adjoint ),
K ( v , v ′ ) : = λ V ( v ∗ v ′ ) = ⟨ Ω , ( V ( v ) ) ∗ V ( v ′ ) Ω ⟩ = ⟨ ξ v , ξ v ′ ⟩ ( v , v ′ ∈ F ) . K(v,v'):=\lambda_{V}(v^{*}v')=\langle\Omega,(V^{(v)})^{*}V^{(v')}\Omega\rangle=\langle\xi_{v},\xi_{v'}\rangle\qquad(v,v'\in F). K ( v , v ′ ) := λ V ( v ∗ v ′ ) = ⟨ Ω , ( V ( v ) ) ∗ V ( v ′ ) Ω ⟩ = ⟨ ξ v , ξ v ′ ⟩ ( v , v ′ ∈ F ) .
Hence K ( v ′ , v ) = K ( v , v ′ ) ‾ K(v',v)=\overline{K(v,v')} K ( v ′ , v ) = K ( v , v ′ ) by the conjugate symmetry of the inner product, and for z : F → C z:F\to\mathbb{C} z : F → C , with η = ∑ v ∈ F z ( v ) ξ v \eta=\sum_{v\in F}z(v)\xi_{v} η = ∑ v ∈ F z ( v ) ξ v , the sesquilinearity of the inner product (conjugate-linear in the first and linear in the second argument, Complex Hilbert Spaces and Bounded Linear Maps: Standing Notation §spaces ) gives ∑ ( v , v ′ ) ∈ F × F z ( v ) ‾ z ( v ′ ) K ( v , v ′ ) = ⟨ η , η ⟩ = ∥ η ∥ 2 \sum_{(v,v')\in F\times F}\overline{z(v)}\,z(v')\,K(v,v')=\langle\eta,\eta\rangle=\lVert\eta\rVert^{2} ∑ ( v , v ′ ) ∈ F × F z ( v ) z ( v ′ ) K ( v , v ′ ) = ⟨ η , η ⟩ = ∥ η ∥ 2 , a nonnegative real number. So K K K is a positive semidefinite kernel on F F F . Thus λ V ∈ L d \lambda_{V}\in\mathcal{L}_{d} λ V ∈ L d .
Proof of clause 2 (realisation). Fix λ ∈ L d \lambda\in\mathcal{L}_{d} λ ∈ L d . We use the notation of Noncommutative Laws, Couplings and the Wasserstein Distance: Standing Notation : P n \mathcal{P}_{n} P n , monomials x v x_{v} x v , Σ n , R \Sigma_{n,R} Σ n , R , Σ n \Sigma_{n} Σ n , and, for a law, its complex GNS space, vacuum vector, classes p ^ \widehat{p} p , multiplication operators L p L_{p} L p and tracial algebra (Noncommutative Laws, Couplings and the Wasserstein Distance: Standing Notation §polynomials , Noncommutative Laws, Couplings and the Wasserstein Distance: Standing Notation §laws , Noncommutative Laws, Couplings and the Wasserstein Distance: Standing Notation §gns ). The monomials of P 2 d \mathcal{P}_{2d} P 2 d are indexed by the words in the letters 1 , … , 2 d 1,\dots,2d 1 , … , 2 d , which are the elements of W 2 d W_{2d} W 2 d : the words and their concatenation are defined by identical clauses in Words over a Finite Alphabet: the Empty Word, Concatenation and Reversal , used for W 2 d W_{2d} W 2 d , and in the earlier version of the definition of words, used by the polynomial items, so this monomial indexing applies; so a variable index v ∈ [ 2 d ] v\in[2d] v ∈ [ 2 d ] is also a letter and has a generator g ( v ) g(v) g ( v ) and a sign ε ( v ) \varepsilon(v) ε ( v ) . Every p ∈ P n p\in\mathcal{P}_{n} p ∈ P n satisfies
p = ∑ v ∈ supp p p ( v ) x v ( p ≠ 0 ) , (E) p=\sum_{v\in\operatorname{supp}p}p(v)\,x_{v}\qquad(p\neq0),\tag{E} p = v ∈ supp p ∑ p ( v ) x v ( p = 0 ) , ( E )
by the formula of Noncommutative Polynomials Form a Unital Complex Algebra with Involution: Linear Extension from Monomials, Products, Adjoints and Self-Adjoint Parts §linear-extension (b) applied to the identity map of P n \mathcal{P}_{n} P n , which is linear with value x v x_{v} x v at x v x_{v} x v ; and x v = x v 1 x v 2 ⋯ x v k x_{v}=x_{v_{1}}x_{v_{2}}\cdots x_{v_{k}} x v = x v 1 x v 2 ⋯ x v k for v v v of length k k k , by Noncommutative Polynomials Form a Unital Complex Algebra with Involution: Linear Extension from Monomials, Products, Adjoints and Self-Adjoint Parts §monomials . The set { 1 , − 1 } k \{1,-1\}^{k} { 1 , − 1 } k of k k k -tuples of signs has 2 k 2^{k} 2 k elements (induction on k k k ), and for maps f m : { 1 , − 1 } → C f_{m}:\{1,-1\}\to\mathbb{C} f m : { 1 , − 1 } → C (m ∈ [ k ] m\in[k] m ∈ [ k ] ), multiplying out by distributivity (induction on k k k ) gives
∑ e ∈ { 1 , − 1 } k ∏ m = 1 k f m ( e m ) = ∏ m = 1 k ( f m ( 1 ) + f m ( − 1 ) ) ; (D) \sum_{e\in\{1,-1\}^{k}}\ \prod_{m=1}^{k}f_{m}(e_{m})=\prod_{m=1}^{k}\bigl(f_{m}(1)+f_{m}(-1)\bigr);\tag{D} e ∈ { 1 , − 1 } k ∑ m = 1 ∏ k f m ( e m ) = m = 1 ∏ k ( f m ( 1 ) + f m ( − 1 ) ) ; ( D )
the same multiplying out holds for products of k k k sums of two terms in P n \mathcal{P}_{n} P n , by the distributive and scalar rules of Noncommutative Polynomials Form a Unital Complex Algebra with Involution: Linear Extension from Monomials, Products, Adjoints and Self-Adjoint Parts §algebra .
Step 1 (the polynomials b l b_{l} b l and P ω P_{\omega} P ω ). For a letter l ∈ [ 2 d ] l\in[2d] l ∈ [ 2 d ] put b l = x g ( l ) + i ε ( l ) x d + g ( l ) ∈ P 2 d b_{l}=x_{g(l)}+i\,\varepsilon(l)\,x_{d+g(l)}\in\mathcal{P}_{2d} b l = x g ( l ) + i ε ( l ) x d + g ( l ) ∈ P 2 d , and for ω ∈ W 2 d \omega\in W_{2d} ω ∈ W 2 d let P ω = b ω P_{\omega}=b_{\omega} P ω = b ω be the product along ω \omega ω of the 2 d 2d 2 d -tuple b = ( b 1 , … , b 2 d ) b=(b_{1},\dots,b_{2d}) b = ( b 1 , … , b 2 d ) ; thus P ∅ = 1 P_{\varnothing}=1 P ∅ = 1 , P ω = b ω 1 ⋯ b ω k P_{\omega}=b_{\omega_{1}}\cdots b_{\omega_{k}} P ω = b ω 1 ⋯ b ω k for ω \omega ω of length k k k , and P u v = P u P v P_{uv}=P_{u}P_{v} P uv = P u P v by Substitution is the Unique Unital Homomorphism with Prescribed Values on the Variables: Monomials, Products, Adjoints and Composition §values . By Noncommutative Polynomials Form a Unital Complex Algebra with Involution: Linear Extension from Monomials, Products, Adjoints and Self-Adjoint Parts §adjoint , x j ∗ = x j x_{j}^{*}=x_{j} x j ∗ = x j and i ε ( l ) ‾ = − i ε ( l ) \overline{i\varepsilon(l)}=-i\varepsilon(l) i ε ( l ) = − i ε ( l ) , so b l ∗ = x g ( l ) − i ε ( l ) x d + g ( l ) = b l − 1 b_{l}^{*}=x_{g(l)}-i\varepsilon(l)x_{d+g(l)}=b_{l^{-1}} b l ∗ = x g ( l ) − i ε ( l ) x d + g ( l ) = b l − 1 , since g ( l − 1 ) = g ( l ) g(l^{-1})=g(l) g ( l − 1 ) = g ( l ) and ε ( l − 1 ) = − ε ( l ) \varepsilon(l^{-1})=-\varepsilon(l) ε ( l − 1 ) = − ε ( l ) . Hence P ω ∗ = P ω ∗ P_{\omega}^{*}=P_{\omega^{*}} P ω ∗ = P ω ∗ for every ω \omega ω : this holds for ω = ∅ \omega=\varnothing ω = ∅ (1 ∗ = 1 1^{*}=1 1 ∗ = 1 ) and for letters, and if it holds for ω \omega ω then ( P ω l ) ∗ = ( P ω b l ) ∗ = b l − 1 P ω ∗ = P l − 1 ω ∗ = P ( ω l ) ∗ (P_{\omega l})^{*}=(P_{\omega}b_{l})^{*}=b_{l^{-1}}P_{\omega^{*}}=P_{l^{-1}\omega^{*}}=P_{(\omega l)^{*}} ( P ω l ) ∗ = ( P ω b l ) ∗ = b l − 1 P ω ∗ = P l − 1 ω ∗ = P ( ω l ) ∗ by ( p q ) ∗ = q ∗ p ∗ (pq)^{*}=q^{*}p^{*} ( pq ) ∗ = q ∗ p ∗ (Noncommutative Polynomials Form a Unital Complex Algebra with Involution: Linear Extension from Monomials, Products, Adjoints and Self-Adjoint Parts §adjoint ); induct on the length.
For a variable index v ∈ [ 2 d ] v\in[2d] v ∈ [ 2 d ] and s ∈ { 1 , − 1 } s\in\{1,-1\} s ∈ { 1 , − 1 } put c ( v , s ) = 1 2 c(v,s)=\tfrac12 c ( v , s ) = 2 1 if ε ( v ) = 1 \varepsilon(v)=1 ε ( v ) = 1 and c ( v , s ) = s 2 i c(v,s)=\tfrac{s}{2i} c ( v , s ) = 2 i s if ε ( v ) = − 1 \varepsilon(v)=-1 ε ( v ) = − 1 . Then
x v = c ( v , 1 ) b ℓ ( g ( v ) , 1 ) + c ( v , − 1 ) b ℓ ( g ( v ) , − 1 ) . (1) x_{v}=c(v,1)\,b_{\ell(g(v),1)}+c(v,-1)\,b_{\ell(g(v),-1)}.\tag{1} x v = c ( v , 1 ) b ℓ ( g ( v ) , 1 ) + c ( v , − 1 ) b ℓ ( g ( v ) , − 1 ) . ( 1 )
Indeed, with j = g ( v ) j=g(v) j = g ( v ) , b ℓ ( j , 1 ) = x j + i x d + j b_{\ell(j,1)}=x_{j}+ix_{d+j} b ℓ ( j , 1 ) = x j + i x d + j and b ℓ ( j , − 1 ) = x j − i x d + j b_{\ell(j,-1)}=x_{j}-ix_{d+j} b ℓ ( j , − 1 ) = x j − i x d + j ; if ε ( v ) = 1 \varepsilon(v)=1 ε ( v ) = 1 then v = j v=j v = j and the right side is 1 2 ( 2 x j ) = x j \tfrac12(2x_{j})=x_{j} 2 1 ( 2 x j ) = x j , and if ε ( v ) = − 1 \varepsilon(v)=-1 ε ( v ) = − 1 then v = d + j v=d+j v = d + j and the right side is 1 2 i ( 2 i x d + j ) = x d + j \tfrac{1}{2i}(2i\,x_{d+j})=x_{d+j} 2 i 1 ( 2 i x d + j ) = x d + j . For a word v v v of length k k k and s ∈ { 1 , − 1 } k s\in\{1,-1\}^{k} s ∈ { 1 , − 1 } k put c ( v , s ) = ∏ m = 1 k c ( v m , s m ) c(v,s)=\prod_{m=1}^{k}c(v_{m},s_{m}) c ( v , s ) = ∏ m = 1 k c ( v m , s m ) and let ω ( v , s ) ∈ W 2 d \omega(v,s)\in W_{2d} ω ( v , s ) ∈ W 2 d be the word of length k k k with m m m -th letter ℓ ( g ( v m ) , s m ) \ell(g(v_{m}),s_{m}) ℓ ( g ( v m ) , s m ) ; for v = ∅ v=\varnothing v = ∅ put c ( ∅ , ⋅ ) = 1 c(\varnothing,\cdot)=1 c ( ∅ , ⋅ ) = 1 and ω ( ∅ , ⋅ ) = ∅ \omega(\varnothing,\cdot)=\varnothing ω ( ∅ , ⋅ ) = ∅ , with a single (empty) sign tuple. Multiplying out x v = x v 1 ⋯ x v k x_{v}=x_{v_{1}}\cdots x_{v_{k}} x v = x v 1 ⋯ x v k with (1) gives
x v = ∑ s ∈ { 1 , − 1 } k c ( v , s ) P ω ( v , s ) . (2) x_{v}=\sum_{s\in\{1,-1\}^{k}}c(v,s)\,P_{\omega(v,s)}.\tag{2} x v = s ∈ { 1 , − 1 } k ∑ c ( v , s ) P ω ( v , s ) . ( 2 )
Consequently, by (E), every p ∈ P 2 d p\in\mathcal{P}_{2d} p ∈ P 2 d with p ≠ 0 p\ne0 p = 0 is a finite combination p = ∑ a ∈ G z ( a ) P ω a p=\sum_{a\in G}z(a)P_{\omega_{a}} p = ∑ a ∈ G z ( a ) P ω a , where G G G is the nonempty finite set of pairs a = ( v , s ) a=(v,s) a = ( v , s ) with v ∈ supp p v\in\operatorname{supp}p v ∈ supp p and s s s a sign tuple of the length of v v v , z ( a ) = p ( v ) c ( v , s ) z(a)=p(v)c(v,s) z ( a ) = p ( v ) c ( v , s ) and ω a = ω ( v , s ) \omega_{a}=\omega(v,s) ω a = ω ( v , s ) (Sums over Finite Index Sets: Finite Unions, Disjoint Unions, Vanishing Terms, Dependent Pairs, Conjugation and the Modulus §pairs ).
Step 2 (the law α \alpha α ). By Noncommutative Polynomials Form a Unital Complex Algebra with Involution: Linear Extension from Monomials, Products, Adjoints and Self-Adjoint Parts §linear-extension (a) there is exactly one linear map α : P 2 d → C \alpha:\mathcal{P}_{2d}\to\mathbb{C} α : P 2 d → C with α ( 1 ) = λ ( ∅ ) = 1 \alpha(1)=\lambda(\varnothing)=1 α ( 1 ) = λ ( ∅ ) = 1 and
α ( x v ) = ∑ s ∈ { 1 , − 1 } k c ( v , s ) λ ( ω ( v , s ) ) for every word v of length k ∈ N . \alpha(x_{v})=\sum_{s\in\{1,-1\}^{k}}c(v,s)\,\lambda\bigl(\omega(v,s)\bigr)\qquad\text{for every word }v\text{ of length }k\in\mathbb{N}. α ( x v ) = s ∈ { 1 , − 1 } k ∑ c ( v , s ) λ ( ω ( v , s ) ) for every word v of length k ∈ N .
Claim A: α ( P ω ) = λ ( ω ) \alpha(P_{\omega})=\lambda(\omega) α ( P ω ) = λ ( ω ) for every ω ∈ W 2 d \omega\in W_{2d} ω ∈ W 2 d . For ω = ∅ \omega=\varnothing ω = ∅ this is α ( 1 ) = 1 = λ ( ∅ ) \alpha(1)=1=\lambda(\varnothing) α ( 1 ) = 1 = λ ( ∅ ) (Laws of d-Tuples of Unitaries §normalised ). Let ω \omega ω have length k k k and j m = g ( ω m ) j_{m}=g(\omega_{m}) j m = g ( ω m ) . Since x ℓ ( j , 1 ) = x j x_{\ell(j,1)}=x_{j} x ℓ ( j , 1 ) = x j and x ℓ ( j , − 1 ) = x d + j x_{\ell(j,-1)}=x_{d+j} x ℓ ( j , − 1 ) = x d + j , we have b ω m = ∑ e ∈ { 1 , − 1 } β m ( e ) x ℓ ( j m , e ) b_{\omega_{m}}=\sum_{e\in\{1,-1\}}\beta_{m}(e)\,x_{\ell(j_{m},e)} b ω m = ∑ e ∈ { 1 , − 1 } β m ( e ) x ℓ ( j m , e ) with β m ( 1 ) = 1 \beta_{m}(1)=1 β m ( 1 ) = 1 and β m ( − 1 ) = i ε ( ω m ) \beta_{m}(-1)=i\varepsilon(\omega_{m}) β m ( − 1 ) = i ε ( ω m ) ; multiplying out,
P ω = ∑ e ∈ { 1 , − 1 } k ( ∏ m = 1 k β m ( e m ) ) x v ( e ) , P_{\omega}=\sum_{e\in\{1,-1\}^{k}}\Bigl(\prod_{m=1}^{k}\beta_{m}(e_{m})\Bigr)x_{v(e)}, P ω = e ∈ { 1 , − 1 } k ∑ ( m = 1 ∏ k β m ( e m ) ) x v ( e ) ,
where v ( e ) v(e) v ( e ) is the word with m m m -th letter ℓ ( j m , e m ) \ell(j_{m},e_{m}) ℓ ( j m , e m ) . Now g ( v ( e ) m ) = j m g(v(e)_{m})=j_{m} g ( v ( e ) m ) = j m and ε ( v ( e ) m ) = e m \varepsilon(v(e)_{m})=e_{m} ε ( v ( e ) m ) = e m , so ω ( v ( e ) , s ) \omega(v(e),s) ω ( v ( e ) , s ) is the word ω s \omega_{s} ω s with m m m -th letter ℓ ( j m , s m ) \ell(j_{m},s_{m}) ℓ ( j m , s m ) , independent of e e e , and c ( v ( e ) m , s m ) c(v(e)_{m},s_{m}) c ( v ( e ) m , s m ) is 1 2 \tfrac12 2 1 if e m = 1 e_{m}=1 e m = 1 and s m 2 i \tfrac{s_{m}}{2i} 2 i s m if e m = − 1 e_{m}=-1 e m = − 1 . By linearity of α \alpha α , exchanging the two finite sums, and (D),
α ( P ω ) = ∑ s ∈ { 1 , − 1 } k λ ( ω s ) ∏ m = 1 k ( 1 2 + i ε ( ω m ) s m 2 i ) = ∑ s ∈ { 1 , − 1 } k λ ( ω s ) ∏ m = 1 k 1 + ε ( ω m ) s m 2 . \alpha(P_{\omega})=\sum_{s\in\{1,-1\}^{k}}\lambda(\omega_{s})\prod_{m=1}^{k}\Bigl(\tfrac12+i\varepsilon(\omega_{m})\tfrac{s_{m}}{2i}\Bigr)=\sum_{s\in\{1,-1\}^{k}}\lambda(\omega_{s})\prod_{m=1}^{k}\tfrac{1+\varepsilon(\omega_{m})s_{m}}{2}. α ( P ω ) = s ∈ { 1 , − 1 } k ∑ λ ( ω s ) m = 1 ∏ k ( 2 1 + i ε ( ω m ) 2 i s m ) = s ∈ { 1 , − 1 } k ∑ λ ( ω s ) m = 1 ∏ k 2 1 + ε ( ω m ) s m .
Each factor 1 + ε ( ω m ) s m 2 \tfrac{1+\varepsilon(\omega_{m})s_{m}}{2} 2 1 + ε ( ω m ) s m is 1 1 1 if s m = ε ( ω m ) s_{m}=\varepsilon(\omega_{m}) s m = ε ( ω m ) and 0 0 0 otherwise. So only the tuple s = ( ε ( ω 1 ) , … , ε ( ω k ) ) s=(\varepsilon(\omega_{1}),\dots,\varepsilon(\omega_{k})) s = ( ε ( ω 1 ) , … , ε ( ω k )) contributes, with product 1 1 1 (Sums over Finite Index Sets: Finite Unions, Disjoint Unions, Vanishing Terms, Dependent Pairs, Conjugation and the Modulus §vanishing ), and for it ω s = ω \omega_{s}=\omega ω s = ω because ℓ ( g ( l ) , ε ( l ) ) = l \ell(g(l),\varepsilon(l))=l ℓ ( g ( l ) , ε ( l )) = l . Hence α ( P ω ) = λ ( ω ) \alpha(P_{\omega})=\lambda(\omega) α ( P ω ) = λ ( ω ) .
Claim B: α ∈ Σ 2 d , 2 \alpha\in\Sigma_{2d,2} α ∈ Σ 2 d , 2 . α \alpha α is linear and α ( 1 ) = 1 \alpha(1)=1 α ( 1 ) = 1 . Let p , q ∈ P 2 d p,q\in\mathcal{P}_{2d} p , q ∈ P 2 d ; if p = 0 p=0 p = 0 or q = 0 q=0 q = 0 then p q = q p = 0 pq=qp=0 pq = qp = 0 and p ∗ p = 0 p^{*}p=0 p ∗ p = 0 (Noncommutative Polynomials Form a Unital Complex Algebra with Involution: Linear Extension from Monomials, Products, Adjoints and Self-Adjoint Parts §algebra ), so the conditions below hold trivially. Otherwise write p = ∑ a ∈ G z ( a ) P ω a p=\sum_{a\in G}z(a)P_{\omega_{a}} p = ∑ a ∈ G z ( a ) P ω a and q = ∑ b ∈ G ′ z ′ ( b ) P ω b ′ q=\sum_{b\in G'}z'(b)P_{\omega'_{b}} q = ∑ b ∈ G ′ z ′ ( b ) P ω b ′ as in Step 1. By Noncommutative Polynomials Form a Unital Complex Algebra with Involution: Linear Extension from Monomials, Products, Adjoints and Self-Adjoint Parts §algebra , P u P v = P u v P_{u}P_{v}=P_{uv} P u P v = P uv , Claim A and Laws of d-Tuples of Unitaries §cyclic ,
α ( p q ) = ∑ ( a , b ) ∈ G × G ′ z ( a ) z ′ ( b ) λ ( ω a ω b ′ ) = ∑ ( a , b ) ∈ G × G ′ z ( a ) z ′ ( b ) λ ( ω b ′ ω a ) = α ( q p ) . \alpha(pq)=\sum_{(a,b)\in G\times G'}z(a)z'(b)\lambda(\omega_{a}\omega'_{b})=\sum_{(a,b)\in G\times G'}z(a)z'(b)\lambda(\omega'_{b}\omega_{a})=\alpha(qp). α ( pq ) = ( a , b ) ∈ G × G ′ ∑ z ( a ) z ′ ( b ) λ ( ω a ω b ′ ) = ( a , b ) ∈ G × G ′ ∑ z ( a ) z ′ ( b ) λ ( ω b ′ ω a ) = α ( qp ) .
By Noncommutative Polynomials Form a Unital Complex Algebra with Involution: Linear Extension from Monomials, Products, Adjoints and Self-Adjoint Parts §adjoint and P ω ∗ = P ω ∗ P_{\omega}^{*}=P_{\omega^{*}} P ω ∗ = P ω ∗ , p ∗ = ∑ a ∈ G z ( a ) ‾ P ω a ∗ p^{*}=\sum_{a\in G}\overline{z(a)}P_{\omega_{a}^{*}} p ∗ = ∑ a ∈ G z ( a ) P ω a ∗ , so by Claim A α ( p ∗ p ) = ∑ ( a , b ) ∈ G × G z ( a ) ‾ z ( b ) λ ( ω a ∗ ω b ) \alpha(p^{*}p)=\sum_{(a,b)\in G\times G}\overline{z(a)}z(b)\lambda(\omega_{a}^{*}\omega_{b}) α ( p ∗ p ) = ∑ ( a , b ) ∈ G × G z ( a ) z ( b ) λ ( ω a ∗ ω b ) . Let F = { ω a : a ∈ G } F=\{\omega_{a}:a\in G\} F = { ω a : a ∈ G } , nonempty and finite, and Z ( ω ) = ∑ a ∈ G , ω a = ω z ( a ) Z(\omega)=\sum_{a\in G,\ \omega_{a}=\omega}z(a) Z ( ω ) = ∑ a ∈ G , ω a = ω z ( a ) for ω ∈ F \omega\in F ω ∈ F . Grouping the terms according to ( ω a , ω b ) (\omega_{a},\omega_{b}) ( ω a , ω b ) , that is, reindexing the finite sum over G × G G\times G G × G as the iterated sum over ( ω , ω ′ ) ∈ F × F (\omega,\omega')\in F\times F ( ω , ω ′ ) ∈ F × F and over the nonempty set of pairs ( a , b ) ∈ G × G (a,b)\in G\times G ( a , b ) ∈ G × G with ω a = ω \omega_{a}=\omega ω a = ω and ω b = ω ′ \omega_{b}=\omega' ω b = ω ′ (Sums over Finite Index Sets: Finite Unions, Disjoint Unions, Vanishing Terms, Dependent Pairs, Conjugation and the Modulus §pairs , together with reindexing along a bijection, claim 2 of Properties of a Sum over a Finite Index Set ), and using distributivity and Sums over Finite Index Sets: Finite Unions, Disjoint Unions, Vanishing Terms, Dependent Pairs, Conjugation and the Modulus §conjugate , gives α ( p ∗ p ) = ∑ ( ω , ω ′ ) ∈ F × F Z ( ω ) ‾ Z ( ω ′ ) λ ( ω ∗ ω ′ ) \alpha(p^{*}p)=\sum_{(\omega,\omega')\in F\times F}\overline{Z(\omega)}Z(\omega')\lambda(\omega^{*}\omega') α ( p ∗ p ) = ∑ ( ω , ω ′ ) ∈ F × F Z ( ω ) Z ( ω ′ ) λ ( ω ∗ ω ′ ) , which is real and nonnegative by Laws of d-Tuples of Unitaries §positive and Positive Semidefinite Kernel on a Finite Set §kernel . So α \alpha α is a tracial state on P 2 d \mathcal{P}_{2d} P 2 d . For v v v of length k k k , ∣ c ( v , s ) ∣ = ( 1 2 ) k |c(v,s)|=(\tfrac12)^{k} ∣ c ( v , s ) ∣ = ( 2 1 ) k by claim 4 of Properties of Complex Conjugation and Modulus (as ∣ s 2 i ∣ = 1 2 |\tfrac{s}{2i}|=\tfrac12 ∣ 2 i s ∣ = 2 1 ), and ∣ λ ( ω ( v , s ) ) ∣ ≤ 1 |\lambda(\omega(v,s))|\le1 ∣ λ ( ω ( v , s )) ∣ ≤ 1 by Unitary Laws: Adjoints, the Bound One, and Reduction to Cyclically Reduced Words §bound ; by Sums over Finite Index Sets: Finite Unions, Disjoint Unions, Vanishing Terms, Dependent Pairs, Conjugation and the Modulus §modulus and the count 2 k 2^{k} 2 k of sign tuples, ∣ α ( x v ) ∣ ≤ 2 k ( 1 2 ) k = 1 ≤ 2 k |\alpha(x_{v})|\le2^{k}(\tfrac12)^{k}=1\le2^{k} ∣ α ( x v ) ∣ ≤ 2 k ( 2 1 ) k = 1 ≤ 2 k . So α \alpha α has norm bound 2 2 2 , i.e. α ∈ Σ 2 d , 2 ⊆ Σ 2 d \alpha\in\Sigma_{2d,2}\subseteq\Sigma_{2d} α ∈ Σ 2 d , 2 ⊆ Σ 2 d .
Step 3 (the joint law γ \gamma γ ). Let s c d \mathrm{sc}_{d} sc d be the semicircular law of d d d variables. It satisfies the Schwinger-Dyson equation, so it equals the vacuum law λ S \lambda_{S} λ S of The Vacuum Law of the Semicircular Operators is the Unique Noncommutative Law Satisfying the Schwinger-Dyson Equation §unique , which belongs to Σ d , 2 \Sigma_{d,2} Σ d , 2 by The Vacuum Law of the Semicircular Operators is the Unique Noncommutative Law Satisfying the Schwinger-Dyson Equation §vacuum-law . Let γ = α ⋆ s c d ∈ Σ 3 d \gamma=\alpha\star\mathrm{sc}_{d}\in\Sigma_{3d} γ = α ⋆ sc d ∈ Σ 3 d be the free product for the split of 3 d = 2 d + d 3d=2d+d 3 d = 2 d + d variables into the first 2 d 2d 2 d and the last d d d , and let ι 1 : P 2 d → P 3 d \iota^{1}:\mathcal{P}_{2d}\to\mathcal{P}_{3d} ι 1 : P 2 d → P 3 d be the substitution of ( x 1 , … , x 2 d ) (x_{1},\dots,x_{2d}) ( x 1 , … , x 2 d ) (Freeness of Two Groups of Variables under a Noncommutative Law §free ). Then γ ∘ ι 1 = α \gamma\circ\iota^{1}=\alpha γ ∘ ι 1 = α by The Free Product of Two Noncommutative Laws: Existence, Uniqueness, Norm Bound, Substitutions and Weak-Star Continuity §product , γ ∈ Σ 3 d , 2 \gamma\in\Sigma_{3d,2} γ ∈ Σ 3 d , 2 by The Free Product of Two Noncommutative Laws: Existence, Uniqueness, Norm Bound, Substitutions and Weak-Star Continuity §bound with R = 2 R=2 R = 2 , and, by Free Semicircular Variables are Characterised by the Schwinger-Dyson Equation Relative to the Other Variables; Semicircular Systems and Rotations §characterisation (with m = 2 d m=2d m = 2 d , n = d n=d n = d ) and The Schwinger-Dyson Equation for a Noncommutative Law §relative ,
γ ( x 2 d + j q ) = ∂ 2 d + j γ ( q ) ( j ∈ [ d ] , q ∈ P 3 d ) , (SD) \gamma(x_{2d+j}\,q)=\partial^{\gamma}_{2d+j}(q)\qquad(j\in[d],\ q\in\mathcal{P}_{3d}),\tag{SD} γ ( x 2 d + j q ) = ∂ 2 d + j γ ( q ) ( j ∈ [ d ] , q ∈ P 3 d ) , ( SD )
with ∂ 2 d + j γ \partial^{\gamma}_{2d+j} ∂ 2 d + j γ the free difference quotient . By Substitution is the Unique Unital Homomorphism with Prescribed Values on the Variables: Monomials, Products, Adjoints and Composition §values , Substitution is the Unique Unital Homomorphism with Prescribed Values on the Variables: Monomials, Products, Adjoints and Composition §homomorphism and Noncommutative Polynomials Form a Unital Complex Algebra with Involution: Linear Extension from Monomials, Products, Adjoints and Self-Adjoint Parts §monomials , ι 1 \iota^{1} ι 1 is linear and multiplicative with ι 1 ( 1 ) = 1 \iota^{1}(1)=1 ι 1 ( 1 ) = 1 and ι 1 ( x v ) = x v \iota^{1}(x_{v})=x_{v} ι 1 ( x v ) = x v for every v ∈ W 2 d v\in W_{2d} v ∈ W 2 d , the word v v v being read as a word in the letters 1 , … , 3 d 1,\dots,3d 1 , … , 3 d all of whose letters are at most 2 d 2d 2 d ; and ι 1 ( p ∗ ) = ι 1 ( p ) ∗ \iota^{1}(p^{*})=\iota^{1}(p)^{*} ι 1 ( p ∗ ) = ι 1 ( p ) ∗ by Substitution is the Unique Unital Homomorphism with Prescribed Values on the Variables: Monomials, Products, Adjoints and Composition §adjoint , the variables being self-adjoint. Put y l = ι 1 ( b l ) = x g ( l ) + i ε ( l ) x d + g ( l ) ∈ P 3 d y_{l}=\iota^{1}(b_{l})=x_{g(l)}+i\varepsilon(l)x_{d+g(l)}\in\mathcal{P}_{3d} y l = ι 1 ( b l ) = x g ( l ) + i ε ( l ) x d + g ( l ) ∈ P 3 d and Q ω = ι 1 ( P ω ) Q_{\omega}=\iota^{1}(P_{\omega}) Q ω = ι 1 ( P ω ) ; thus Q ∅ = 1 Q_{\varnothing}=1 Q ∅ = 1 , Q ω = y ω 1 ⋯ y ω k Q_{\omega}=y_{\omega_{1}}\cdots y_{\omega_{k}} Q ω = y ω 1 ⋯ y ω k , Q u Q v = Q u v Q_{u}Q_{v}=Q_{uv} Q u Q v = Q uv , Q ω ∗ = Q ω ∗ Q_{\omega}^{*}=Q_{\omega^{*}} Q ω ∗ = Q ω ∗ , and, by Claim A,
γ ( Q ω ) = α ( P ω ) = λ ( ω ) ( ω ∈ W 2 d ) . (3) \gamma(Q_{\omega})=\alpha(P_{\omega})=\lambda(\omega)\qquad(\omega\in W_{2d}).\tag{3} γ ( Q ω ) = α ( P ω ) = λ ( ω ) ( ω ∈ W 2 d ) . ( 3 )
Step 4 (the operators). Let ( H , M , Ω ) = ( H γ , M γ , Ω γ ) (H,M,\Omega)=(\mathcal{H}_{\gamma},\mathcal{M}_{\gamma},\Omega_{\gamma}) ( H , M , Ω ) = ( H γ , M γ , Ω γ ) , a tracial W*-probability space by The Tracial Algebra of a Noncommutative Law is a Tracial W*-Probability Space: the W*-Closure of the Left Multiplications §w-star , with M γ = A γ ′ ′ \mathcal{M}_{\gamma}=\mathcal{A}_{\gamma}'' M γ = A γ ′′ where A γ = { L p : p ∈ P 3 d } \mathcal{A}_{\gamma}=\{L_{p}:p\in\mathcal{P}_{3d}\} A γ = { L p : p ∈ P 3 d } . Every L p L_{p} L p belongs to M M M : it commutes with every element of A γ ′ \mathcal{A}_{\gamma}' A γ ′ by The Commutant of a Set of Bounded Operators on a Complex Hilbert Space §commutant , so L p ∈ A γ ′ ′ L_{p}\in\mathcal{A}_{\gamma}'' L p ∈ A γ ′′ . The trace is τ M ( T ) = ⟨ Ω , T Ω ⟩ \tau_{M}(T)=\langle\Omega,T\Omega\rangle τ M ( T ) = ⟨ Ω , T Ω ⟩ (Tracial W*-Probability Spaces §trace , Cyclic Tracial Operator Algebras and Their Traces §trace ). Since γ ∈ Σ 3 d , 2 \gamma\in\Sigma_{3d,2} γ ∈ Σ 3 d , 2 , Left and Right Multiplication Operators on the Complex GNS Space of a Noncommutative Law: Boundedness, Algebra Rules, Adjoints, Commutation, the Vacuum and the Conjugation applies with r = 2 r=2 r = 2 : for p , q ∈ P 3 d p,q\in\mathcal{P}_{3d} p , q ∈ P 3 d and c ∈ C c\in\mathbb{C} c ∈ C , L p + q = L p + L q L_{p+q}=L_{p}+L_{q} L p + q = L p + L q , L c p = c L p L_{cp}=cL_{p} L c p = c L p , L p q = L p L q L_{pq}=L_{p}L_{q} L pq = L p L q , L 1 = I L_{1}=I L 1 = I (Left and Right Multiplication Operators on the Complex GNS Space of a Noncommutative Law: Boundedness, Algebra Rules, Adjoints, Commutation, the Vacuum and the Conjugation §algebra ), L p ∗ = L p ∗ L_{p}^{*}=L_{p^{*}} L p ∗ = L p ∗ (Left and Right Multiplication Operators on the Complex GNS Space of a Noncommutative Law: Boundedness, Algebra Rules, Adjoints, Commutation, the Vacuum and the Conjugation §adjoint ), ∥ L x j ∥ o p ≤ 2 \lVert L_{x_{j}}\rVert_{\mathrm{op}}\le2 ∥ L x j ∥ op ≤ 2 (Left and Right Multiplication Operators on the Complex GNS Space of a Noncommutative Law: Boundedness, Algebra Rules, Adjoints, Commutation, the Vacuum and the Conjugation §multiplication ), and L p Ω = p ^ L_{p}\Omega=\widehat{p} L p Ω = p , ⟨ p ^ , q ^ ⟩ = γ ( p ∗ q ) \langle\widehat{p},\widehat{q}\rangle=\gamma(p^{*}q) ⟨ p , q ⟩ = γ ( p ∗ q ) , τ M ( L p ) = γ ( p ) \tau_{M}(L_{p})=\gamma(p) τ M ( L p ) = γ ( p ) (Left and Right Multiplication Operators on the Complex GNS Space of a Noncommutative Law: Boundedness, Algebra Rules, Adjoints, Commutation, the Vacuum and the Conjugation §vacuum ).
For j ∈ [ d ] j\in[d] j ∈ [ d ] put U j = L y j U_{j}=L_{y_{j}} U j = L y j , where y j = x j + i x d + j y_{j}=x_{j}+ix_{d+j} y j = x j + i x d + j , and S j = L x 2 d + j S_{j}=L_{x_{2d+j}} S j = L x 2 d + j ; both belong to M M M . Each S j S_{j} S j is self-adjoint (Left and Right Multiplication Operators on the Complex GNS Space of a Noncommutative Law: Boundedness, Algebra Rules, Adjoints, Commutation, the Vacuum and the Conjugation §adjoint , x 2 d + j x_{2d+j} x 2 d + j being self-adjoint) with ∥ S j ∥ o p ≤ 2 \lVert S_{j}\rVert_{\mathrm{op}}\le2 ∥ S j ∥ op ≤ 2 . For a letter l l l we have U ( l ) = L y l U^{(l)}=L_{y_{l}} U ( l ) = L y l in the notation of the statement: if ε ( l ) = 1 \varepsilon(l)=1 ε ( l ) = 1 then l = g ( l ) l=g(l) l = g ( l ) and this is the definition; if ε ( l ) = − 1 \varepsilon(l)=-1 ε ( l ) = − 1 then l = d + j l=d+j l = d + j with j = g ( l ) j=g(l) j = g ( l ) and U ( l ) = U j ∗ = L y j ∗ U^{(l)}=U_{j}^{*}=L_{y_{j}^{*}} U ( l ) = U j ∗ = L y j ∗ with y j ∗ = x j − i x d + j = y d + j y_{j}^{*}=x_{j}-ix_{d+j}=y_{d+j} y j ∗ = x j − i x d + j = y d + j . By L p q = L p L q L_{pq}=L_{p}L_{q} L pq = L p L q and L 1 = I L_{1}=I L 1 = I it follows that
U ( ω ) = L Q ω ( ω ∈ W 2 d ) . (4) U^{(\omega)}=L_{Q_{\omega}}\qquad(\omega\in W_{2d}).\tag{4} U ( ω ) = L Q ω ( ω ∈ W 2 d ) . ( 4 )
Unitarity. Let l l l be a letter and r = Q l l − 1 − 1 r=Q_{l\,l^{-1}}-1 r = Q l l − 1 − 1 . Since ( l l − 1 ) ∗ = ( l − 1 ) − 1 l − 1 = l l − 1 (l\,l^{-1})^{*}=(l^{-1})^{-1}l^{-1}=l\,l^{-1} ( l l − 1 ) ∗ = ( l − 1 ) − 1 l − 1 = l l − 1 , we get r ∗ = r r^{*}=r r ∗ = r and r ∗ r = Q l l − 1 l l − 1 − 2 Q l l − 1 + 1 r^{*}r=Q_{l\,l^{-1}\,l\,l^{-1}}-2Q_{l\,l^{-1}}+1 r ∗ r = Q l l − 1 l l − 1 − 2 Q l l − 1 + 1 (Noncommutative Polynomials Form a Unital Complex Algebra with Involution: Linear Extension from Monomials, Products, Adjoints and Self-Adjoint Parts §algebra ). By (3) and Laws of d-Tuples of Unitaries §cancellation (with u = ∅ u=\varnothing u = ∅ and v = l l − 1 v=l\,l^{-1} v = l l − 1 , resp. u = v = ∅ u=v=\varnothing u = v = ∅ ) and Laws of d-Tuples of Unitaries §normalised , γ ( Q l l − 1 l l − 1 ) = λ ( l l − 1 ) = λ ( ∅ ) = 1 \gamma(Q_{l\,l^{-1}\,l\,l^{-1}})=\lambda(l\,l^{-1})=\lambda(\varnothing)=1 γ ( Q l l − 1 l l − 1 ) = λ ( l l − 1 ) = λ ( ∅ ) = 1 , so ∥ L r Ω ∥ 2 = ⟨ r ^ , r ^ ⟩ = γ ( r ∗ r ) = 1 − 2 + 1 = 0 \lVert L_{r}\Omega\rVert^{2}=\langle\widehat{r},\widehat{r}\rangle=\gamma(r^{*}r)=1-2+1=0 ∥ L r Ω ∥ 2 = ⟨ r , r ⟩ = γ ( r ∗ r ) = 1 − 2 + 1 = 0 . As L r ∈ M L_{r}\in M L r ∈ M and ( H , M , Ω ) (H,M,\Omega) ( H , M , Ω ) is a cyclic tracial operator algebra, L r = 0 L_{r}=0 L r = 0 by The Trace, the Conjugation and the Right Action of a Cyclic Tracial Operator Algebra §separating . By (4), U ( l ) U ( l − 1 ) = L Q l l − 1 = L r + L 1 = I U^{(l)}U^{(l^{-1})}=L_{Q_{l\,l^{-1}}}=L_{r}+L_{1}=I U ( l ) U ( l − 1 ) = L Q l l − 1 = L r + L 1 = I . For l = j l=j l = j this reads U j U j ∗ = I U_{j}U_{j}^{*}=I U j U j ∗ = I , and for l = d + j l=d+j l = d + j it reads U j ∗ U j = I U_{j}^{*}U_{j}=I U j ∗ U j = I . By Fact U, every U j U_{j} U j is unitary. By (4) and (3), λ U ( ω ) = τ M ( L Q ω ) = γ ( Q ω ) = λ ( ω ) \lambda_{U}(\omega)=\tau_{M}(L_{Q_{\omega}})=\gamma(Q_{\omega})=\lambda(\omega) λ U ( ω ) = τ M ( L Q ω ) = γ ( Q ω ) = λ ( ω ) for every ω \omega ω , i.e. λ U = λ \lambda_{U}=\lambda λ U = λ .
Elements of A U \mathcal{A}_{U} A U . Let Z ∈ A U Z\in\mathcal{A}_{U} Z ∈ A U , say Z = ∑ r = 1 N c r U ( ω r ) Z=\sum_{r=1}^{N}c_{r}U^{(\omega_{r})} Z = ∑ r = 1 N c r U ( ω r ) . By (4) and the algebra rules, Z = L q Z=L_{q} Z = L q with q = ∑ r c r Q ω r = ι 1 ( q ~ ) q=\sum_{r}c_{r}Q_{\omega_{r}}=\iota^{1}(\tilde q) q = ∑ r c r Q ω r = ι 1 ( q ~ ) , q ~ = ∑ r c r P ω r ∈ P 2 d \tilde q=\sum_{r}c_{r}P_{\omega_{r}}\in\mathcal{P}_{2d} q ~ = ∑ r c r P ω r ∈ P 2 d . If q ~ ≠ 0 \tilde q\neq0 q ~ = 0 , then by (E) and the properties of ι 1 \iota^{1} ι 1 ,
q = ∑ v ∈ supp q ~ q ~ ( v ) x v , τ M ( Z ) = γ ( q ) = ∑ v ∈ supp q ~ q ~ ( v ) γ ( x v ) , (5) q=\sum_{v\in\operatorname{supp}\tilde q}\tilde q(v)\,x_{v},\qquad\tau_{M}(Z)=\gamma(q)=\sum_{v\in\operatorname{supp}\tilde q}\tilde q(v)\,\gamma(x_{v}),\tag{5} q = v ∈ supp q ~ ∑ q ~ ( v ) x v , τ M ( Z ) = γ ( q ) = v ∈ supp q ~ ∑ q ~ ( v ) γ ( x v ) , ( 5 )
where every v v v is a word all of whose letters are at most 2 d 2d 2 d ; if q ~ = 0 \tilde q=0 q ~ = 0 then Z = L 0 = 0 Z=L_{0}=0 Z = L 0 = 0 .
(a) Centred. Let i ∈ [ d ] i\in[d] i ∈ [ d ] and Z = L q Z=L_{q} Z = L q as above. Then τ M ( S i Z ) = τ M ( L x 2 d + i q ) = γ ( x 2 d + i q ) = ∂ 2 d + i γ ( q ) \tau_{M}(S_{i}Z)=\tau_{M}(L_{x_{2d+i}q})=\gamma(x_{2d+i}q)=\partial^{\gamma}_{2d+i}(q) τ M ( S i Z ) = τ M ( L x 2 d + i q ) = γ ( x 2 d + i q ) = ∂ 2 d + i γ ( q ) by (SD). If q ~ = 0 \tilde q=0 q ~ = 0 this is 0 0 0 . Otherwise, by (5) and linearity, it is ∑ v q ~ ( v ) ∂ 2 d + i γ ( x v ) \sum_{v}\tilde q(v)\,\partial^{\gamma}_{2d+i}(x_{v}) ∑ v q ~ ( v ) ∂ 2 d + i γ ( x v ) ; here ∂ 2 d + i γ ( x ∅ ) = ∂ 2 d + i γ ( 1 ) = 0 \partial^{\gamma}_{2d+i}(x_{\varnothing})=\partial^{\gamma}_{2d+i}(1)=0 ∂ 2 d + i γ ( x ∅ ) = ∂ 2 d + i γ ( 1 ) = 0 , and for v v v of length k k k every term of the sum defining ∂ 2 d + i γ ( x v ) \partial^{\gamma}_{2d+i}(x_{v}) ∂ 2 d + i γ ( x v ) in The Free Difference Quotient and the Conjugate Variables of a Noncommutative Law §difference-quotient carries the factor δ v m , 2 d + i = 0 \delta_{v_{m},2d+i}=0 δ v m , 2 d + i = 0 , since v m ≤ 2 d < 2 d + i v_{m}\le2d<2d+i v m ≤ 2 d < 2 d + i . So τ M ( S i Z ) = 0 \tau_{M}(S_{i}Z)=0 τ M ( S i Z ) = 0 .
(b) Covariance. Let i , l ∈ [ d ] i,l\in[d] i , l ∈ [ d ] and Z = L q Z=L_{q} Z = L q , Z ′ = L q ′ Z'=L_{q'} Z ′ = L q ′ as above. Then τ M ( S i Z S l Z ′ ) = γ ( x 2 d + i q x 2 d + l q ′ ) = ∂ 2 d + i γ ( q x 2 d + l q ′ ) \tau_{M}(S_{i}ZS_{l}Z')=\gamma(x_{2d+i}\,q\,x_{2d+l}\,q')=\partial^{\gamma}_{2d+i}(q\,x_{2d+l}\,q') τ M ( S i Z S l Z ′ ) = γ ( x 2 d + i q x 2 d + l q ′ ) = ∂ 2 d + i γ ( q x 2 d + l q ′ ) by (SD). If q ~ = 0 \tilde q=0 q ~ = 0 or q ~ ′ = 0 \tilde q'=0 q ~ ′ = 0 , both this and δ i l τ M ( Z ) τ M ( Z ′ ) \delta_{il}\tau_{M}(Z)\tau_{M}(Z') δ i l τ M ( Z ) τ M ( Z ′ ) vanish. Otherwise, by (5), the algebra rules and x v x 2 d + l x v ′ = x v ( 2 d + l ) v ′ x_{v}x_{2d+l}x_{v'}=x_{v\,(2d+l)\,v'} x v x 2 d + l x v ′ = x v ( 2 d + l ) v ′ (Noncommutative Polynomials Form a Unital Complex Algebra with Involution: Linear Extension from Monomials, Products, Adjoints and Self-Adjoint Parts §monomials ),
∂ 2 d + i γ ( q x 2 d + l q ′ ) = ∑ v ∈ supp q ~ ∑ v ′ ∈ supp q ~ ′ q ~ ( v ) q ~ ′ ( v ′ ) ∂ 2 d + i γ ( x v ( 2 d + l ) v ′ ) . \partial^{\gamma}_{2d+i}(q\,x_{2d+l}\,q')=\sum_{v\in\operatorname{supp}\tilde q}\ \sum_{v'\in\operatorname{supp}\tilde q'}\tilde q(v)\tilde q'(v')\,\partial^{\gamma}_{2d+i}\bigl(x_{v\,(2d+l)\,v'}\bigr). ∂ 2 d + i γ ( q x 2 d + l q ′ ) = v ∈ supp q ~ ∑ v ′ ∈ supp q ~ ′ ∑ q ~ ( v ) q ~ ′ ( v ′ ) ∂ 2 d + i γ ( x v ( 2 d + l ) v ′ ) .
The word w = v ( 2 d + l ) v ′ w=v\,(2d+l)\,v' w = v ( 2 d + l ) v ′ has the letter 2 d + l 2d+l 2 d + l at the position n 0 n_{0} n 0 following the letters of v v v , while all its other letters are at most 2 d 2d 2 d and so differ from 2 d + i 2d+i 2 d + i . Hence in The Free Difference Quotient and the Conjugate Variables of a Noncommutative Law §difference-quotient only the term at n 0 n_{0} n 0 can be nonzero; it carries δ 2 d + l , 2 d + i = δ i l \delta_{2d+l,2d+i}=\delta_{il} δ 2 d + l , 2 d + i = δ i l , and there w < n 0 = v w_{<n_{0}}=v w < n 0 = v and w > n 0 = v ′ w_{>n_{0}}=v' w > n 0 = v ′ . So ∂ 2 d + i γ ( x v ( 2 d + l ) v ′ ) = δ i l γ ( x v ) γ ( x v ′ ) \partial^{\gamma}_{2d+i}(x_{v(2d+l)v'})=\delta_{il}\gamma(x_{v})\gamma(x_{v'}) ∂ 2 d + i γ ( x v ( 2 d + l ) v ′ ) = δ i l γ ( x v ) γ ( x v ′ ) , and summing with (5),
τ M ( S i Z S l Z ′ ) = δ i l ( ∑ v q ~ ( v ) γ ( x v ) ) ( ∑ v ′ q ~ ′ ( v ′ ) γ ( x v ′ ) ) = δ i l τ M ( Z ) τ M ( Z ′ ) . \tau_{M}(S_{i}ZS_{l}Z')=\delta_{il}\Bigl(\sum_{v}\tilde q(v)\gamma(x_{v})\Bigr)\Bigl(\sum_{v'}\tilde q'(v')\gamma(x_{v'})\Bigr)=\delta_{il}\,\tau_{M}(Z)\,\tau_{M}(Z'). τ M ( S i Z S l Z ′ ) = δ i l ( v ∑ q ~ ( v ) γ ( x v ) ) ( v ′ ∑ q ~ ′ ( v ′ ) γ ( x v ′ ) ) = δ i l τ M ( Z ) τ M ( Z ′ ) .
This proves clause 2 with ( H , M , Ω ) (H,M,\Omega) ( H , M , Ω ) , U = ( U 1 , … , U d ) U=(U_{1},\dots,U_{d}) U = ( U 1 , … , U d ) and S 1 , … , S d S_{1},\dots,S_{d} S 1 , … , S d . ■ \blacksquare ■