Abstract
We consider functions of pairs of noncommuting contractions on Hilbert space and study the problem as to which functions we have Lipschitz type estimates in Schatten–von Neumann norms. We prove that if belongs to the Besov class of analytic functions in the bidisc, then we have a Lipschitz type estimate for functions of pairs of not necessarily commuting contractions in the Schatten–von Neumann norms for . On the other hand, we show that for functions in , there are no such Lipschitz type estimates for , nor in the operator norm.
Export citation and abstract BibTeX RIS
The research of the first author was partially supported by RFBR grant no. 17-01-00607A. The publication was prepared with the support of the RUDN University Program "5-100". |
§ 1. Introduction
The purpose of this paper is to study the behaviour of functions of (not necessarily commuting) contractions and under perturbation. We are going to obtain Lipschitz type estimates in the Schatten–von Neumann norms , , for functions in the Besov class of analytic functions. Note that functions of noncommuting contractions can be defined in terms of double operator integrals with respect to semi-spectral measures, see §3 below.
This paper can be considered as a continuation of the results of [1]–[18] for functions of perturbed self-adjoint operators, contractions, normal operators, dissipative operators, functions of collections of commuting self-adjoint operators and functions of collections of noncommuting self-adjoint operators.
Recall that a Lipschitz function on does not have to be operator Lipschitz, i.e., the condition , , does not imply that
for arbitrary self-adjoint operators (bounded or unbounded, it does not matter) and . This was first established in [19].
It turned out that functions in the (homogeneous) Besov space are operator Lipschitz; this was established in [1] and [3] (see [20] for detailed information about Besov classes). We refer the reader to the recent survey [12] for detailed information on operator Lipschitz functions. In particular, [12] presents various sufficient conditions and necessary conditions for a function on to be operator Lipschitz. It is well known that if is an operator Lipschitz function on , and and are self-adjoint operators such that the difference belongs to the Schatten–von Neumann class , , then and . Moreover, the constant on the right does not depend on . In particular, this is true for functions in the Besov class , i.e.,
However, it was discovered in [9] (see also [21]) that the situation becomes quite different if we replace the class of Lipschitz functions with the class of Hölder functions of order , . Namely, the inequality , , implies that
for arbitrary self-adjoint operators and . Moreover, it was shown in [10] that if , , and , then and
for arbitrary self-adjoint operators and .
Analogs of the above results for functions of normal operators, functions of contractions, functions of dissipative operators and functions of commuting collections of self-adjoint operators were obtained in [2], [11], [14] and [15].
Note that it was shown in [17] that for , inequality (1.1) holds for arbitrary Lipschitz (not necessarily operator Lipschitz) functions with constant on the right that depends on . An analog of this result for functions of commuting self-adjoint operators was obtained in [18].
In [16] similar problems were considered for functions of two noncommuting self-adjoint operators (such functions can be defined in terms of double operator integrals, see [16]). It was shown in [16] that for functions on in the (homogeneous) Besov class and for the following Lipschitz type estimate holds:
for arbitrary pairs and of (not necessarily commuting) self-adjoint operators.
However, it was shown in [16] that for there is no such Lipschitz type estimate in the norm, nor in the operator norm. Moreover, it follows from the construction given in [16] that for and for positive numbers , , there exists a function in with Fourier transform supported in such that and
while
Here we use the notation for the operator norm.
This implies that, unlike in the case of commuting operators, there cannot be any Hölder type estimates in the operator norm for Hölder functions of order . Moreover, for , there can be no estimate for for functions in the Besov class for any and .
On the other hand, it was observed by the anonymous referee of [16] that, unlike in the case of commuting self-adjoint operators, there are no Lipschitz type estimates for for Lipschitz functions on , see [16].
Finally, let us mention that in the case of functions of triples of noncommuting self-adjoint operators there are no such Lipschitz type estimates for functions in the Besov class in the norm of for any . This was established in [8].
In §3 we give an introduction to double and triple operator integrals and we define functions of noncommuting contractions. We also define the Haagerup and Haagerup-like tensor products of three copies of the disc-algebra and we define triple operator integrals whose integrands belong to such tensor products.
Lipschitz type estimates in the Schatten–von Neumann norm will be obtained in §4. We show that for and for any function on in the analytic Besov space the following Lipschitz type inequality holds:
for arbitrary pairs and of contractions. Recall that a similar inequality was established in [16] for functions of self-adjoint operators. However, to obtain this inequality for functions of contractions, we need new algebraic formulae. Moreover, to obtain this inequality for functions of contractions, we offer an approach that does not use triple operator integrals. To be more precise, we reduce the inequality to the case of analytic polynomials . Then we have to integrate over finite sets, in which case triple operator integrals become finite sums. We establish explicit representations of the operator differences for analytic polynomials in terms of finite sums of elementary tensors which allows us to estimate the norms.
However, we still use triple operator integrals to obtain in §5 explicit formulae for the operator differences for arbitrary functions in .
In §6 we study differentiability properties in Schatten–von Neumann norms of the function
for and contractive-valued functions and . We obtain explicit formulae for the derivative in terms of triple operator integrals. Again, to prove the existence of the derivative, we do not need triple operator integrals.
As in the case of functions of pairs of noncommuting self-adjoint operators (see [16]), there are no Lipschitz type estimates in the norm of , , for functions of pairs of not necessarily commuting contractions , . This will be established in §7. Note that the construction given in §7 differs from the construction in the case of self-adjoint operators given in [16].
In §8 we state some open problems and in §2 we give an introduction to Besov classes on polydiscs.
Throughout this paper we deal with Schatten–von Neumann ideals . We refer the reader to the book [22] for detailed information about these classes.
We use the notation for the normalized Lebesgue measure on the unit circle and the notation for the normalized Lebesgue measure on . For simplicity, we assume that we deal with separable Hilbert spaces.
§ 2. Besov classes of periodic functions
In this section we give a brief introduction to Besov spaces on the torus.
To define Besov spaces on the torus , we consider an infinitely differentiable function on such that
Let , , be the trigonometric polynomials defined by
where
For a distribution on , we put
It is easy to see that
the series converges in the sense of distributions. We say that belongs the Besov class , , , if
The analytic subspace of consists of functions in for which the Fourier coefficients satisfy the equalities:
We refer the reader to [20] for more detailed information about Besov spaces.
§ 3. Double and triple operator integrals with respect to semi-spectral measures
In this section we give a brief introduction to double and triple operator integrals with respect to semi-spectral measures.
3.1. Double operator integrals
Double operator integrals with respect to spectral measures are expressions of the form
where and are spectral measures, is a linear operator and is a bounded measurable function. They appeared first in [23]. Later Birman and Solomyak developed in [24]–[26] a beautiful theory of double operator integrals.
Double operator integrals with respect to semi-spectral measures were defined in [2], see also [12] (recall that the definition of a semi-spectral measure differs from the definition of a spectral measure by the fact that the condition that it takes values in the set of orthogonal projections is replaced with the condition that it takes values in the set of non-negative contractions, see [12] for more detail).
For the double operator integral to make sense for an arbitrary bounded linear operator , we have to impose an additional assumption on . The natural class of such functions is called the class of Schur multipliers, see [1]. There are various characterizations of the class of Schur multipliers, see [1] and [27]. In particular, is a Schur multiplier if and only if it belongs to the Haagerup tensor product of and , i.e., it admits a representation of the form
where the and satisfy the condition
In this case
the series converges in the weak operator topology. The right-hand side of this equality does not depend on the choice of a representation of in (3.2).
One can also consider double operator integrals of the form (3.1) in the case when and are semi-spectral measures. In this case, as in the case of spectral measures, formula (3.4) still holds under the same assumption (3.3).
It is easy to see that if belongs to the projective tensor product of and , i.e., admits a representation of the form (3.2) with and satisfying
then is a Schur multiplier and (3.4) holds.
3.2. The semi-spectral measures of contractions
Recall that if is a contraction (i.e., ) on a Hilbert space , then by the Sz.-Nagy dilation theorem (see [28]), has a unitary dilation, i.e., there exist a Hilbert space that contains and a unitary operator on such that
where is the orthogonal projection onto .
Among all unitary dilations of one can always select a minimal unitary dilation (in a natural sense) and all minimal unitary dilations are isomorphic, see [28].
The existence of a unitary dilation allows us to construct the natural functional calculus for functions in the disc-algebra defined by
where is the spectral measure of .
Consider the operator set function defined on the Borel subsets of the unit circle by
Then is a semi-spectral measure. It can be shown that it does not depend on the choice of a unitary dilation. The semi-spectral measure is called the semi-spectral measure of .
3.3. Functions of noncommuting contractions
Let be a function on the torus that belongs to the Haagerup tensor product , i.e., admits a representation of the form
where , are functions in such that
For a pair of (not necessarily commuting) contractions, the operator is defined as the double operator integral
Note that if , then , and so we can take functions of contractions for an arbitrary function in . Indeed, if is an analytic polynomial in two variables of degree at most in each variable, then we can represent in the form
Thus belongs to the projective tensor product and
It follows easily from (2.3) that every function of Besov class belongs to , and so the operator is well defined by formula (3.5). Clearly,
where is the polynomial defined by (2.1). It follows immediately from (3.6) and (2.3) that the series converges absolutely in the operator norm. Note that formula (3.7) can be used as a definition of the functions of noncommuting contractions in the case when .
3.4. Triple operator integrals. Haagerup tensor products
There are several approaches to multiple operator integrals. Triple operator integrals are expressions of the form
where is a bounded measurable function, , and are spectral measures, and and are bounded linear operators on Hilbert space.
In [5] triple (and more general, multiple) operator integrals were defined for functions in the integral projective product . For such functions , the following Schatten–von Neumann properties hold:
whenever . Later in [29] triple (and multiple) operator integrals were defined for functions in the Haagerup tensor product . However, it turns out that under the assumption the conditions and imply that
only under the conditions that and , see [30] (see also [16]). Moreover, the following inequality holds:
whenever and , see [30].
Note also that, to obtain Lipschitz type estimates for functions of noncommuting self-adjoint operators in [16], we had to use triple operator integrals with integrands that do not belong to the Haagerup tensor product . That is why we had to introduce in [16] Haagerup-like tensor products of the first kind and of the second kind.
In this paper we are going to use triple operator integrals whose integrands are continuous functions on that belong to Haagerup and Haagerup-like tensor products of three copies of the disc-algebra . We briefly define such tensor products and discuss inequalities we are going to use in the next section.
Definition 1. We say that a continuous function on belongs to the Haagerup tensor product if admits a representation
where , are are functions in such that
Here stands for the operator norm of a matrix (finite or infinite) on the space or on a finite-dimensional Euclidean space. By definition, the norm of in is the infimum of the left-hand side of (3.9) over all representations of in the form of (3.8).
Suppose that and both (3.8) and (3.9) hold. Let , and be contractions with semi-spectral measures , and . Then for bounded linear operators and , we can define the triple operator integral
as
It is easy to verify that the series converges in the weak operator topology if we consider partial sums over rectangles. It can be shown in the same way as in the case of triple operator integrals with respect to spectral measures that the sum on the right does not depend on the choice of a representation of in the form of (3.8), see Theorem 3.1 of [16].
We are going to use Lemma 3.2 of [30]. Suppose that is a sequence of bounded linear operators on Hilbert space such that
Let be a bounded linear operator. Consider the row and the column defined by
and
Then by Lemma 3.2 of [30], for , the following inequalities hold:
whenever . Recall that Lemma 3.2 in [30] is proved first for and , after which the interpolation theorem is applied.
It is easy to verify that under the above assumptions,
where is the operator matrix .
Lemma 3.1. Under the above hypotheses,
Proof. Let be a unitary dilation of the contraction on a Hilbert space , . Clearly, we can consider the space as a subspace of . It is easy to see that
where is the orthogonal projection onto . The result follows from the inequality
which is a consequence of the spectral theorem.
It follows from Lemma 3.2 of [31] that under the above assumptions, inequalities (3.11) hold for , , with
and for , , with
This together with Lemma 3.1 and inequalities (3.12) implies that for and , , , under the above assumptions, the following inequality holds:
where .
The following theorem is an analog of the corresponding result for triple operator integrals with respect to spectral measures, see [30]. It follows immediately from (3.14).
Theorem 3.2. Let , and be contractions, and let and , , . Suppose that . Then , , and
Recall that by we mean the class of bounded linear operators.
3.5. Haagerup-like tensor products
We define here Haagerup-like tensor products of disc-algebras by analogy with Haagerup-like tensor products of spaces, see [16].
Definition 2. A continuous function on is said to belong to the Haagerup-like tensor product of the first kind if it admits a representation
where , and are functions in such that
Clearly, if and only if the function
belongs to the Haagerup tensor product .
Similarly, we can define the Haagerup-like tensor product of the second kind.
Definition 3. A continuous function on is said to belong to the Haagerup-like tensor product of the second kind if it admits a representation
where , are are functions in such that
Let us first consider the situation when is defined by (3.15) or by (3.16) with summation over a finite set. In this case triple operator integrals of the form (3.10) can be defined for arbitrary functions , for arbitrary bounded linear operators and and for arbitrary contractions , and .
Suppose that
where and are finite sets. We put
Suppose now that
where and are finite sets. Then we put
The following estimate is a very special case of Theorem 3.4 below. However, we have stated it separately because its proof is elementary and does not require the definition of triple operator integrals.
Theorem 3.3. Let and be bounded linear operators, let , and be contractions, and let and . Suppose that and are finite sets. The following statements hold:
(i) Let be given by (3.17). Suppose that and . Then the sum on the right of (3.18) belongs to and
(ii) Let be given by (3.19). Suppose that and . Then the sum on the right of (3.20) belongs to and
Proof. Let us prove (i). The proof of (ii) is the same. We are going to use a duality argument. Suppose that and , . We have
3.6. Triple operator integrals with integrands in Haagerup-like tensor products
We define triple operator integrals with respect to semi-spectral measures with integrands in by analogy with triple operator integrals with respect to spectral measures, see [16] and [30]. Let and let . Suppose that , and are contractions. For an operator of class and for a bounded linear operator , we define the triple operator integral
as the following continuous linear functional on , (on the class of compact operators in the case ):
Note that the triple operator integral
is well defined as the integrand belongs to the Haagerup tensor product .
Again, we can define triple operator integrals with integrands in , by analogy with the case of spectral measures, see [16] and [30]. Let and let , and be contractions. Suppose that is a bounded linear operator and , . The triple operator integral
is defined as the continuous linear functional
on (on the class of compact operators if ).
As in the case of spectral measures (see [30]), the following theorem can be proved:
Theorem 3.4. Suppose that , and are contractions, and let and . The following statements hold:
(1) Let . Suppose that and . Then the operator in (3.21) belongs to and
(2) Let . Suppose that and . Then the operator in (3.22) belongs to and
§ 4. Lipschitz type estimates in Schatten–von Neumann norms
In this section we obtain Lipschitz type estimates in the Schatten–von Neumann classes for for functions of contractions. To obtain such estimates, we are going to use an elementary approach and obtain elementary formulae that involve only finite sums.
Later we will need explicit expressions for operator differences, which will be obtained in the next section in terms of triple operator integrals. Such formulae will be used to obtain formulae for operator derivatives.
Suppose that is a function that belongs to the Besov space of analytic functions (see §2). As we have observed in §3.3, we can define functions for (not necessarily commuting) contractions and on Hilbert space by formula (3.7).
For a differentiable function on we use the notation for the divided difference:
For a differentiable function on we define the divided differences and by
We need several elementary identities. Let be the set of th roots of :
and let
The following elementary formulae are well known. We give proofs for completeness.
Lemma 4.1. Let and be analytic polynomials in one variable of degree less than . Then
In particular,
Proof. It suffices to consider the case when and with and note that
In the same way we can obtain similar formulae for polynomials in several variables. We need only the case of two variables.
Lemma 4.3. Let and be polynomials in two variables of degree less than in each variable. Then
In particular,
Proof. It suffices to consider the case when and with . Then and
Suppose now that and are pairs of not necessarily commuting contractions.
Theorem 4.4. Let be an analytic polynomial in two variable of degree at most in each variable. Then
We are going to establish (4.2). The proof of (4.3) is similar. We need the following lemma.
Lemma 4.5. Let be an analytic polynomial in one variable of degree at most . Then
Thus,
if , . Hence,
It follows that
whenever , .
Let . It is easy to see that
Hence,
Proof of Theorem 4.4. Clearly, it suffices to prove (4.2) in the case when , where is a polynomial of one variable of degree at most and . Clearly, in this case
On the other hand,
Identity (4.2) follows now from Lemma 4.5.
For , we denote by the integral operator on with kernel function , i.e.,
The following lemma allows us to evaluate the operator norm of this operator for polynomials of degree less than in each variable in terms of the operator norms of the matrix .
Lemma 4.6. Let be an analytic polynomial in two variables of degree less than in each variable. Then
where and . Hence,
where the supremum is taken over all polynomials and in one variable of degree less than and such that , . Next, by Lemma 4.3, for arbitrary polynomials and with and , we have
It remains to observe that by Lemma 4.1, if and only if
and the same is true for .
Theorem 4.7. Let be a trigonometric polynomial in one variable of degree at most . Then
Proof. The result follows from Lemma 4.6 and the inequality
which is a consequence of the fact that is equal to the norm of the Hankel operator on the Hardy class , see [4], Ch. 1, Theorem 1.10.
Corollary 4.8. Let be a trigonometric polynomial in two variables of degree at most in each variable and let . Suppose that , , , are contractions such that and . Then
This estimate is a consequence of formula (4.2), Theorem 3.3, Theorem 4.7 and Corollary 4.2. The norm can be estimated in a similar way. This implies the desired inequality.
Corollary 4.8 allows us to establish a Lipschitz type inequality for functions in .
Theorem 4.9. Let and let . Suppose that , , , are contractions such that and . Then
Proof. The result follows immediately from Corollary 4.8 and inequality (2.3).
§ 5. A representation of operator differences in terms of triple operator integrals
In this section we obtain an explicit formula for the operator differences , , in terms of triple operator integrals.
Lemma 5.2. Let be an analytic polynomial in two variables of degree at most in each variable. Then
Proof. Both formulae (5.1) and (5.2) can be verified straightforwardly. However, we deduce them from Theorem 4.4.
Formula (5.1) follows immediately from formula (4.2), if we consider the special case when , and are the operators of multiplication by , and on the one-dimensional space. Similarly, formula (5.2) follows immediately from formula (4.3).
Corollary 5.3. Under the hypotheses of Lemma 5.2, the following inequalities hold:
Proof. The result is a consequence of Lemma 5.2, Theorem 4.7, Corollary 4.2 and Definitions 2 and 3.
Proof of Theorem 5.1. The result follows immediately from Corollary 5.3 and inequality (2.3).
Theorem 5.4. Let . Suppose that , , , are contractions such that and . Then for , the following formula holds:
Proof. Suppose first that is an analytic polynomial in two variables of degree at most in each variable. In this case equality (5.3) is a consequence of Theorem 4.4, Lemma 5.2 and the definition of triple operator integrals given in §3.5.
In the general case we represent by the series (2.1) and apply (5.3) to each . The result follows from (2.3).
§ 6. Differentiability properties
In this section we study differentiability properties of the map
in the norm of , , for functions and that take contractive values and are differentiable in .
We say that an operator-valued function defined on an interval is differentiable in if for any , and the limit
exists in the norm of for each in .
Theorem 6.1. Let and let . Suppose that and are operator-valued functions on an interval that take contractive values and are differentiable in . Then the function (6.1) is differentiable on in and
.
Proof. As before, it suffices to prove the result in the case when is an analytic polynomial. Let be a positive integer such that has degree at most in each variable. Put . We have
Clearly,
in the norm of . On the other hand, it is easy to see that
in the operator norm. Hence,
It follows now from Lemma 5.2 and from the definition of triple operator integrals given in §3 that the right-hand side is equal to
which completes the proof.
§ 7. The case
In this section we show that unlike in the case , there are no Lipschitz type estimates in the norm of in the case when for functions , , of noncommuting contractions. In particular, there are no such Lipschitz type estimates for functions in the operator norm. Moreover, we show that for , such Lipschitz type estimates do not hold even for functions and for pairs of noncommuting unitary operators.
Recall that similar results were obtained in [16] for functions of noncommuting self-adjoint operators. However, in this paper we use a different construction to obtain such results for functions of unitary operators.
Lemma 7.1. For each matrix there exists an analytic polynomial in two variables of degree at most in each variable such that for all and .
Clearly, for all and
by Corollary 4.2.
Lemma 7.2. For each , there exist an analytic polynomial in two variables of degree at most in each variable, and unitary operators , and such that
for every .
Proof. One can select orthonormal bases and in an -dimensional Hilbert space such that for all . Indeed, let be the subspace of of analytic polynomials of degree less than . We can put and , where , .
Consider the rank one orthogonal projections and defined by , , and , . We define the unitary operators , , and by
By Lemma 7.1 applied to the set , there exists an analytic polynomial in two variables of degree at most in each variable such that
and . It follows easily from (7.1) that and
We have
Hence, and
It remains to observe that .
Remark. If we replace the polynomial constructed in the proof of Lemma 7.2 with the polynomial defined by
it will obviously satisfy the same inequality:
It is easy to deduce from (2.3) that for such a polynomial
for some constants and .
This together with (7.2) implies the following result:
Theorem 7.3. Let and let . Then there exist unitary operators , , and an analytic polynomial in two variables such that
§ 8. Open problems
In this section we state open problems for functions of noncommuting contractions.
8.1. Functions of triples of contractions
Recall that it was shown in [8] that for there are no Lipschitz type estimates in the norm of for any for functions of triples of noncommuting self-adjoint operators. We conjecture that the same must be true in the case of functions of triples of not necessarily commuting contractions. Note that the construction given in [8] does not generalize to the case of functions of contractions.
8.2. Lipschitz functions of noncommuting contractions
Recall that an anonymous referee of [16] observed that for Lipschitz functions on , there are no Lipschitz type estimates for functions of noncommuting self-adjoint operators in the Hilbert–Schmidt norm. The construction is given in [16]. We conjecture that the same result must hold in the case of functions of noncommuting contractions.
8.3. Lipschitz type estimates for and Hölder type estimates
It follows from results of [16] that in the case of functions of noncommuting self-adjoint operators for any , and , there exist pairs of self-adjoint operators and and a function in the homogeneous Besov space such that can be arbitrarily large while can be arbitrarily small. In particular, the condition does not imply any Lipschitz or Hölder type estimates in the norm of for any positive and .
It is easy to see that in the case of contractions the situation is different: for any and , there exists such that the condition guarantees a Lipschitz type estimate for functions of not necessarily commuting contractions in .
It would be interesting to find optimal conditions on that would guarantee Lipschitz or Hölder type estimates in for a given .