Harmonic Analysis of Operators on Hilbert Space (Universitext)

Yields q(T )n = ρ · pr q(U)n (n = 1, 2, . . . ). (11. eleven) simply because |q(λ )| ≤ 1 for |λ | ≤ 1, it follows from the spectral idea of unitary operators that q(U) ≤ 1. for this reason, there exists a unitary operator V in a few better house such that q(U)n = pr V n (n = zero, 1, . . . ). (11. 12) Now (11. eleven) and (11. 12) suggest q(T )n = ρ · pr V n (n = 1, 2, . . . ), and accordingly q(T ) ∈ Cρ . For ρ = 2 this consequence is also restated, by means of advantage of Proposition eleven. 2, within the following shape. Proposition eleven. 6. If w(T ) ≤ 1 then w(q(T )) ≤ 1 for each polynomial q(λ ) such that q(0) = zero and |q(λ )| ≤ 1 (|λ | ≤ 1). specifically, w(T ) ≤ 1 implies w(T n ) ≤ 1 (n = 1, 2, . . . ). allow us to nation back that the result of this part relate to operators on complicated Hilbert areas. 12 Notes Theorem 1. 1 at the decomposition of an area H brought about by means of an isometry V on H has been formulated in a probabilistic atmosphere by way of W outdated [1], p. 89. aside from the expression of the subspace H0 of the unitary half because the intersection of the levels of the iterates of V , the theory already appears to be like within the primary paper on summary Hilbert area of VON N EUMANN [1], p. ninety six; in its current shape the concept used to be said and proved through H ALMOS [2], Lemma 1. Proposition 2. 1 on bilateral shifts can be derived—at least for advanced Hilbert space—from the final thought of spectral multiplicity; the direct evidence given the following, that is legitimate with no limit at the box of scalars, is because of H ALPERIN; see Sz. -N. –F. [V]. Proposition three. 1 at the invariant vectors of a contraction used to be discovered through S Z . -NAGY in reference to a few ergodic theorems (cf. R IESZ–S Z. -NAGY [1] and [Func. Anal. ] Sec. 144). Generalizations of this proposition got in S Z. -N. –F. [1]. Theorem three. 2 at the canonical decomposition of a contraction used to be proved via L ANGER [1] and S Z. -N. –F. [IV]. The notation A = pr B used to be brought by means of S Z . -NAGY in [P]. for 2 operators so comparable, H ALMOS [1] says is a “compression” of B, and B is a “dilation” of 50 C HAPTER I. C ONTRACTIONS AND T inheritor D ILATIONS A, while S Z. -NAGY [P] says “projection” rather than “compression”. during this e-book we've most popular to desert this terminology and protect for the time period “dilation” the that means given in Sec. four (i. e. , “power dilation” within the experience of H ALMOS [4]). incidentally, allow us to become aware of that, simply because A = pr B if and provided that the bilinear shape (Bb, b′ ) is an extension of the bilinear shape (Aa, a′ ), we might be justified in calling B a “numerical extension” of A, and A a “numerical limit” of B (in analogy with “numerical range”, “numerical radius”, and so on. ). the truth that for each contraction T there exists an isometry V such that T = pr V , was once saw already through J ULIA [1]–[3]. H ALMOS [1] has proven that V will be selected to be unitary, V = U. Theorem four. 2, at the lifestyles of a unitary U such that the family members T n = pr U n carry at the same time for n = 1, 2, . . . (i. e. , of a unitary dilation of T ), was once stumbled on by way of S Z . -NAGY [I]. the unique evidence used the concept of F. R IESZ at the trigonometric second challenge and the theory of N A˘I MARK [1] (Theorem eight.