Each result cited below is used for the data in its own statement. For ε∈{1,2}, the maps πε and Eε are those of Two Noncommutative Laws with a Common Marginal: Standing Notation for Their Amalgamated Free Product §embeddings. By Marginals of a Noncommutative Law: the Isometry of GNS Spaces, the Trace-Preserving Embedding of Tracial Algebras and the Conditional Expectation §expectation, for S,T∈N and c∈Aε,
Eε(πε(S)cπε(T))=SEε(c)Tandτμ(Eε(c))=τγε(c);
by Marginals of a Noncommutative Law: the Isometry of GNS Spaces, the Trace-Preserving Embedding of Tracial Algebras and the Conditional Expectation §homomorphism, πε(I)=I; by The Tracial Algebra of a Noncommutative Law: a Norm-Closed Unital *-Algebra with a Faithful Positive Trace, Determined by Vacuum Vectors, Closed under Square Roots §trace (for the law γε), τγε(cc′)=τγε(c′c) for c,c′∈Aε; and all products below lie in the algebra indicated by The Tracial Algebra of a Noncommutative Law: a Norm-Closed Unital *-Algebra with a Faithful Positive Trace, Determined by Vacuum Vectors, Closed under Square Roots §star-algebra. Taking S=I or T=I in the first identity gives xEε(c)=Eε(πε(x)c) and Eε(c)z=Eε(cπε(z)) for x,z∈N.
Step 1 (one layer). Let ε∈{1,2}, u,v∈Aε and x,z∈N. Then
τμ(xEε(uπε(z)v))=τμ(Eε(vπε(x)u)z).(1)
Indeed, by the identities above the left side equals τμ(Eε(πε(x)uπε(z)v))=τγε(πε(x)uπε(z)v), which by the trace property (with c=πε(x)uπε(z) and c′=v) equals τγε(vπε(x)uπε(z))=τμ(Eε(vπε(x)uπε(z)))=τμ(Eε(vπε(x)u)z).
Step 2 (induction on k). For k∈N let P(k) be the assertion: for all f1,…,fk∈{1,2}, all uj,vj∈Afj (j∈[k]) and all x,y∈N, the elements Gk and Gk′ defined in the statement satisfy τμ(xGk)=τμ(Gk′y). Let K={k∈N:P(k)}; we show that 1∈K and that k+1∈K whenever k∈K, so that K=N by Principle of Induction for the Natural Numbers.
For k=1: G1=Ef1(u1πf1(y)v1) and G1′=Ef1(v1πf1(x)u1), and P(1) is (1) with ε=f1, u=u1, v=v1, z=y.
Let k∈K and let data f1,…,fk+1, uj,vj, x,y be given, with G1,…,Gk+1 and G1′,…,Gk+1′ as in the statement. The first k elements G1,…,Gk are also the elements G1,…,Gk of the truncated data (fj,uj,vj)j∈[k] with the same y, since both are given by the same first term and the same recursion for the indices j≤k (induction on j). Put
x′=Efk+1(vk+1πfk+1(x)uk+1)∈N.
Since Gk+1=Efk+1(uk+1πfk+1(Gk)vk+1), (1) with ε=fk+1 and z=Gk gives τμ(xGk+1)=τμ(x′Gk). By P(k) for the truncated data, with x′ in place of x and the same y,
τμ(x′Gk)=τμ(Hky),H1=Efk(vkπfk(x′)uk),Hj=Efk+1−j(vk+1−jπfk+1−j(Hj−1)uk+1−j) (2≤j≤k).
For the data of length k+1 the statement defines G1′=x′ and Gj′=Efk+2−j(vk+2−jπfk+2−j(Gj−1′)uk+2−j) for 2≤j≤k+1. We claim Gj+1′=Hj for every j∈[k], by induction on j: for j=1, G2′=Efk(vkπfk(G1′)uk)=Efk(vkπfk(x′)uk)=H1; and if j<k and Gj+1′=Hj, then, since k+2−(j+2)=k+1−(j+1)=k−j,
Gj+2′=Efk−j(vk−jπfk−j(Gj+1′)uk−j)=Efk−j(vk−jπfk−j(Hj)uk−j)=Hj+1.
Hence Gk+1′=Hk, and τμ(xGk+1)=τμ(x′Gk)=τμ(Gk+1′y). This is P(k+1), so k+1∈K.
Therefore P(k) holds for every k∈N, which is the claim.