Volume 2011, Issue 1 472495
Research Article
Open Access

Multiplication Operators between Lipschitz-Type Spaces on a Tree

Robert F. Allen

Robert F. Allen

Department of Mathematics, University of Wisconsin-La Crosse, La Crosse, WI 54601, USA uwlax.edu

Search for more papers by this author
Flavia Colonna

Corresponding Author

Flavia Colonna

Department of Mathematical Sciences, George Mason University, Fairfax, VA 22030, USA gmu.edu

Search for more papers by this author
Glenn R. Easley

Glenn R. Easley

System Planning Corporation, Arlington, VA 22209, USA sysplan.com

Search for more papers by this author
First published: 05 June 2011
Citations: 10
Academic Editor: Ingo Witt

Abstract

Let be the space of complex-valued functions f on the set of vertices T of an infinite tree rooted at o such that the difference of the values of f at neighboring vertices remains bounded throughout the tree, and let w be the set of functions f such that |f(v) − f(v)| = O(|v|−1), where |v| is the distance between o and v and v is the neighbor of v closest to o. In this paper, we characterize the bounded and the compact multiplication operators between and w and provide operator norm and essential norm estimates. Furthermore, we characterize the bounded and compact multiplication operators between w and the space L of bounded functions on T and determine their operator norm and their essential norm. We establish that there are no isometries among the multiplication operators between these spaces.

1. Introduction

Let 𝒳 and 𝒴 be complex Banach spaces of functions defined on a set Ω. For a complex-valued function ψ defined on Ω, the multiplication operator with symbol ψ from 𝒳 to 𝒴 is defined as
(1.1)
A fundamental objective in the study of the operators with symbol is to tie the properties of the operator to the function theoretic properties of the symbol.
When Ω is taken to be the open unit disk 𝔻 in the complex plane, an important space of functions to study is the Bloch space, defined as the set of analytic functions f : 𝔻 for which
(1.2)
The Bloch space can also be described as the set consisting of the Lipschitz functions between metric spaces from 𝔻 endowed with the Poincaré distance ρ to endowed with the Euclidean distance, a fact that was proved by the second author in [1] (see also [2]). In fact, f if and only if there exist β > 0 such that for all z, w𝔻
(1.3)

More recently, considerable research has been carried out in the field of operator theory when the set Ω is taken to be a discrete structure, such as a discrete group or a graph. In this paper, we consider the case when Ω is taken to be an infinite tree.

By a tree T we mean a locally finite, connected, and simplyconnected graph, which, as a set, we identify with the collection of its vertices. Two vertices u and v are called neighbors if there is an edge connecting them, and we use the notation u ~ v. A vertex is called terminal if it has a unique neighbor. A path is a finite or infinite sequence of vertices [v0, v1, …] such that vk ~ vk+1 and vk−1vk+1, for all k.

Given a tree T rooted at o and a vertex uT, a vertex v is called a descendant of u if u lies in the unique path from o to v. The vertex u is then called an ancestor of v. Given a vertex vo, we denote by v the unique neighbor which is an ancestor of v. For vT, the set Sv consisting of v and all its descendants is called the sector determined by v.

Define the length of a finite path [u = u0, u1, …, v = un] (with uk ~ uk+1 for k = 0, …, n) to be the number n of edges connecting u to v. The distance, d(u, v), between vertices u and v is the length of the path connecting u to v. The tree T is a metric space under the distance d. Fixing o as the root of the tree, we define the length of a vertex v by |v | = d(o, v). By a function on a tree we mean a complex-valued function on the set of its vertices.

In this paper, the tree will be assumed to be rooted at a vertex o and without terminal vertices (and hence infinite).

Infinite trees are discrete structures which exhibit significant geometric and potential theoretic characteristics that are present in the Poincaré disk 𝔻. For instance, they have a boundary, which is defined as the set of equivalence classes of paths which differ by finitely many vertices. The union of the boundary with the tree yields a compact space. A useful resource for the potential theory on trees illustrating the commonalities with the disk is [3]. In [4] it was shown that, if the tree has the property that all its vertices have the same number of neighbors, then there is a natural embedding of the tree in the unit disk such that the edges of the tree are arcs of geodesics in 𝔻 with the same hyperbolic length and the set of cluster points of the vertices is the entire unit circle.

In [5], the last two authors defined the Lipschitz space on a tree T as the set consisting of the functions f : T which are Lipschitz with respect to the distance d on T and the Euclidean distance on . For this reason, the Lipschitz space can be viewed as a discrete analogue of the Bloch space . It was also shown that the Lipschitz functions on T are precisely the functions for which
(1.4)
where Df(v) = |f(v) − f(v)| and T* = T∖{o}. Under the norm
(1.5)
is a Banach space containing the space L of the bounded functions on T. Furthermore, for fL, ∥f ≤ 2∥f.
The little Lipschitz space is defined as
(1.6)
and was proven to be a separable closed subspace of . We state the following results that will be useful in the present paper.

Lemma 1.1 (see [5], Lemma  3.4.) (a) If f and vT, then

(1.7)
In particular, if ∥f ≤ 1, then |f(v)|≤|v| for each vT*.

(b) If f0, then

(1.8)

Lemma 1.2 (see [5], Proposition  2.4.)Let {fn} be a sequence of functions in 0 converging to 0 pointwise in T such that {∥fn} is bounded. Then fn → 0 weakly in 0.

In [6], we introduced the weighted Lipschitz space on a tree T as the set w of the functions f : T such that
(1.9)
The interest in this space is due to its connection to the bounded multiplication operators on . Specifically, it was shown in [5] that the bounded multiplication operators on are precisely those operators Mψ whose symbol ψ is a bounded function in w. The space w was shown to be a Banach space under the norm
(1.10)
The little weighted Lipschitz space was defined as
(1.11)
and was shown to be a closed separable subspace of w.

In this paper, we will make repeated use of the following results proved in [6].

Lemma 1.3 (see [6], Propositions  2.1 and  2.6.) (a) If fw, and vT*, then

(1.12)

(b) If fw,0, then

(1.13)

Lemma 1.4 (see [6], Proposition   2.7.)Let {fn} be a sequence of functions in w,0 converging to 0 pointwise in T such that {∥fnw} is bounded. Then fn → 0 weakly in w,0.

In this paper, we consider the multiplication operators between and w, as well as between w and L. The multiplication operators between and L were studied by the last two authors in [7].

1.1. Organization of the Paper

In Sections 2 and 3, we study the multiplication operators between w and . We characterize the bounded and the compact operators and give estimates on their operator norm and their essential norm. We also prove that no isometric multiplication operators exist between the respective spaces.

In Section 4, we characterize the bounded operators and the compact operators from w to L and determine their operator norm and their essential norm. As was the case in Sections 2 and 3, we show that no isometries exist amongst such operators. In addition, we characterize the multiplication operators that are bounded from below.

Finally, in Section 5, we characterize the bounded and the compact multiplication operators from L to w. We also determine their operator norm and their essential norm. As with all the other cases, we show that there are no isometries amongst such operators.

2. Multiplication Operators from w to

We begin the section with the study of the bounded multiplication operators Mψ : w and Mψ : w,00.

2.1. Boundedness and Operator Norm Estimates

Let ψ be a function on the tree T. Define
(2.1)
In the following theorem, we give a boundedness criterion in terms of the quantities τψ and σψ.

Theorem 2.1. For a function ψ on T, the following statements are equivalent:

(a)  Mψ : w is bounded.

(b)  Mψ : w,00 is bounded.

(c)  τψ and σψ are finite.

Furthermore, under these conditions, we have

(2.2)

Proof. (a)(c) Assume Mψ : w is bounded. Applying Mψ to the constant function 1, we have ψ, so that, by Lemma 1.1, we have σψ < . Next, consider the function f on T defined by f(v) = log (1+|v|). Then f(o) = 0; for vo, a straightforward calculation shows that

(2.3)
and lim |v|→ |v | Df(v) = 1. Thus, ∥fw = 1 and so ∥Mψf ≤ ∥Mψ∥. Therefore, for vT*, noting that
(2.4)
one has
(2.5)
Hence τψ < .

(c)(a) Assume τψ and σψ are finite. Then, by Lemma 1.3, for fw and vT*, we have

(2.6)
Since |ψ(o)| ≤ σψ, we obtain
(2.7)
proving the boundedness of Mψ : w and the upper estimate.

(b)(c) Suppose Mψ : w,00 is bounded. The finiteness of σψ follows again from the fact that ψ = Mψ1 ∈ 0 and from Lemma 1.1. To prove that τψ < , let 0 < α < 1 and, for vT, define fα(v) = (log (1+|v|))α. Then fα(o) = 0 and |v | Dfa(v) → 0 as |v | → ; so fαw,0. Since for 0 < α < 1, the function xxxα is increasing for x ≥ 1, the function Dfα(v) is increasing in α, and Dfα(v) ≤ Df(v) for vT*, where f(v) = log (1+|v|), for vT. Thus, ∥fαw ≤ ∥fw = 1. Therefore, for vT*, we have

(2.8)
Letting α → 1, we obtain
(2.9)
Hence τψ < .

(c)(b) Assume σψ and τψ are finite, and let fw,0. Then, by Lemma 1.3, for vT*, we have

(2.10)
as |v | → . Thus, ψf0. The boundedness of Mψ and the estimate ∥Mψfτψ + σψ can be shown as in the proof of (c)(a).

Finally we show that, under boundedness assumptions on Mψ, ∥Mψ∥≥max {τψ, σψ}. For vT*, let fv = 1/(|v | + 1)χv, where χv denotes the characteristic function of {v}. Then ∥fvw = 1 and

(2.11)
Furthermore, letting fo = (1/2)χo, we see that and ∥ψfo = |ψ(o)|. Therefore, we deduce that ∥Mψ∥≥σψ.

Next, fix vT* and for wT, define

(2.12)
Then, gvw and
(2.13)
Observe that, for wT*, we have
(2.14)
Hence
(2.15)
Define fv = gv/∥gvw. Then ∥fvw = 1 and
(2.16)
Taking the limit as |v | → , we obtain ∥Mψ∥≥τψ. Therefore, ∥Mψ∥≥max {τψ, σψ}.

2.2. Isometries

In this section, we show there are no isometric multiplication operators Mψ from the spaces w or w,0 to the spaces or 0, respectively.

Assume Mψ : w is an isometry. Then ∥ψ = ∥Mψ1∥ = 1. On the other hand, . Thus , which implies that ψ is a constant of modulus 1. Yet, for vT*, letting fv = (1/(|v | + 1))χv, we see that
(2.17)
which yields a contradiction. Therefore, we obtain the following result.

Theorem 2.2. There are no isometries Mψ from w to or w,0 to 0, respectively.

2.3. Compactness and Essential Norm Estimates

In this section, we characterize the compact multiplication operators. As with many classical spaces, the characterization of the compact operators is a “little-oh” condition corresponding the the “big-oh” condition for boundedness. We first collect some useful results about compact operators from w or w,0 to .

Lemma 2.3. A bounded multiplication operator Mψ from w to is compact if and only if for every bounded sequence {fn} in w converging to 0 pointwise, the sequence {∥ψfn} → 0 as n.

Proof. Assume Mψ is compact, and let {fn} be a bounded sequence in w converging to 0 pointwise. Without loss of generality, we may assume ∥fnw ≤ 1 for all n. Then the sequence {Mψfn} = {ψfn} has a subsequence which converges in the -norm to some function f. Clearly , and by part (a) of Lemma 1.1, for vT*, we have

(2.18)
Thus, pointwise on T. Since fn → 0 pointwise, it follows that f must be identically 0, which implies that . With 0 being the only limit point of {ψfn} in , it follows that as n.

Conversely, assume every bounded sequence {fn} in w converging to 0 pointwise has the property that as n. Let {gn} be a sequence in w with for all n. Then |gn(o)| ≤ 1 for all n, and by part (a) of Lemma 1.2, for vT*, we obtain

(2.19)
Thus, {gn} is uniformly bounded on finite subsets of T. So some subsequence converges pointwise to some function g. Fix vT* and ɛ > 0. Then for k sufficiently large, we have
(2.20)
We deduce
(2.21)
for all k sufficiently large. So gw. The sequence defined by is bounded in w and converges to 0 pointwise. Thus by hypothesis, we obtain as k. It follows that in the -norm, thus proving the compactness of Mψ.

By an analogous argument, we obtain the corresponding compactness criterion for Mψ from w,0 to 0.

Lemma 2.4. A bounded multiplication operator Mψ from w,0 to 0 is compact if and only if for every bounded sequence {fn} in w,0 converging to 0 pointwise, the sequence {∥ψfn} → 0 as n.

The following result is a variant of Lemma 1.3(a), which will be needed to prove a characterization of the compact multiplication operators from w to and from w,0 to 0 (Theorem 2.6).

Lemma 2.5. For fw and  vT

(2.22)
where .

Proof. Fix vT and argue by induction on n = |v|. For n = 0, inequality (2.22) is obvious. So assume |v | = n > 0 and |f(u)|≤|f(o)| + 2log (1+|u|)sw(f) for all vertices u such that |u | < n. Then

(2.23)
Next, observe that 1/(|v | + 1)≤log ((|v | + 1)/|v|), so
(2.24)
Hence
(2.25)
Inequality (2.22) now follows immediately from (2.23) and (2.25).

Theorem 2.6. Let Mψ be a bounded multiplication operator from w to (or equivalently from w,0 to 0). Then the following statements are equivalent:

(a)  Mψ : w is compact.

(b)  Mψ : w,00 is compact.

(c)  lim |v|→ | ψ(v)|/(|v | + 1) = 0 and lim |v|→Dψ(v)log  | v | = 0.

Proof. We first prove (a)(c). Assume Mψ : w is compact. It suffices to show that, for any sequence {vn} in T such that 2≤|vn | → , we have lim n | ψ(vn)|/(|vn | + 1) = 0 and lim nDψ(vn)log  | vn | = 0. Let {vn} be such a sequence, and for each n, define . Then fn(o) = 0, fn → 0 pointwise as n, and ∥fnw = 1. By Lemma 2.3, it follows that ∥ψfn → 0 as n. Furthermore

(2.26)
Thus lim n |ψ(vn)|/(|vn | + 1) = 0.

Next, for each n and vT, define

(2.27)
Then Dgn(v) = 0   if or |v | >|vn | − 1. In addition, if , then |v | Dgn(v) < 4. Indeed, there are two possibilities. Either , in which case
(2.28)
or , in which case
(2.29)
Thus {∥gnw} is bounded, and {gn} converges to 0 pointwise. By Lemma 2.3, it follows that ∥ψgn → 0 as n. Moreover
(2.30)
Therefore lim n  Dψ(vn)log |vn| = 0.

To prove the implication (c)(a), suppose lim |v|→Dψ(v)log  | v | = 0 and lim |v|→ | ψ(v)|/(|v | + 1) = 0. Clearly, if ψ is identically 0, then Mψ is compact. So assume Mψ : w is bounded with ψ not identically 0. By Lemma 2.3, it suffices to show that if {fn} is bounded in w converging to 0 pointwise, then ∥ψfn → 0 as n. Let {fn} be such a sequence, , and fix ɛ > 0. Note that

(2.31)
Thus there exists an M such that
(2.32)
for |v | ≥ M. Using Lemma 2.5, for |v| > M, we have
(2.33)

On the other hand, on the set BM = {vT:|v | ≤ M}, {fn} converges to 0 uniformly, and thus Dfn does as well. Moreover

(2.34)
uniformly on BM. Therefore D(ψfn) → 0 uniformly on T. Furthermore, the sequence {(ψfn)(o)} converges to 0 as n. Hence ∥ψfn → 0 as n, proving that Mψ is compact.

Finally, note that the functions fn and gn defined in the proof of (a)(c) are in w,0. So the equivalence of (b) and (c) is proved analogously.

Recall the essential norm of a bounded operator S between Banach spaces 𝒳 and 𝒴 is defined as
(2.35)
For ψ a function on T, define the quantities
(2.36)

Theorem 2.7. Let Mψ be a bounded multiplication operator from w to . Then

(2.37)

Proof. For each n, define fn = (1/(n + 1))χn, where χn denotes the characteristic function of the set {vT:|v | = n}. Then fnw,0, ∥fnw = 1, and fn → 0 pointwise. Thus, by Lemma 1.4, {fn} converges to 0 weakly in w,0. Let 𝒦 be the set of compact operators from w,0 to 0, and let K𝒦. Then K is completely continuous [8], and so ∥Kfn → 0 as n. Thus

(2.38)
Now note that
(2.39)
Hence
(2.40)
We will now show that . This estimate is clearly true if B(ψ) = 0. So assume {vn} is a sequence in T such that 2≤|vn | → as n and
(2.41)
For n and vT, define
(2.42)
Then hn(o) = 0, , and
(2.43)

The supremum of |v | Dhn(v) is attained at the vertices of length |vn | − 1 and is given by

(2.44)
Since (|vn| − 1)log (|vn|/(|vn| − 1)) ≤ 1, we have
(2.45)

By letting , we have gnw,0, ∥gnw = 1, and gn → 0 pointwise. By Lemma 1.4, the sequence {gn} converges to 0 weakly in w,0. Thus ∥Kgn → 0 as n. Therefore,

(2.46)

For each n, we have . So

(2.47)
Since lim nsn = 1, we have
(2.48)
Therefore, ∥Mψe ≥ max {A(ψ), B(ψ)}.

We now derive an upper estimate on the essential norm.

Theorem 2.8. Let Mψ be a bounded multiplication operator from w to . Then

(2.49)

Proof. For n, define the operator Kn on w by

(2.50)
where fw and vn is the ancestor of v of length n. For fw, (Knf)(o) = f(o), and Knfw,0. Let Bn = {vT : |v| ≤ n}, and note that Knf attains finitely many values, whose number does not exceed the cardinality of Bn. Let {gk} be a sequence in w such that ∥gkw ≤ 1 for each k. Then a = sup k | gk(o)| ≤ 1, and |Kngk(o)| ≤ a. Furthermore, by part (a) of Lemma 1.3, for each vT* and for each k, we have |Kngk(v)| ≤ 1 + log n. Thus, some subsequence of {Kngk} k must converge to a function g on T attaining constant values on the sectors determined by the vertices of length n. It follows that this subsequence converges to g in w as well, proving that Kn is a compact operator on w. Since Mψ is bounded as an operator from w to , it follows that MψKn : w is compact for all n.

Define the operator Jn = IKn, where I denotes the identity operator on w. Then Jnf(o) = 0, and for vT*, we have

(2.51)
By part (a) of Lemma 1.3, we see that
(2.52)

Using (2.51) and (2.52), we obtain

(2.53)
Since
(2.54)
taking the limit as n, we obtain
(2.55)
as desired.

3. Multiplication Operators from to w

We begin this section with a boundedness criterion for the multiplication operators from Mψ : w and Mψ : 0w,0.

3.1. Boundedness and Operator Norm Estimates

Let ψ be a function on the tree T. Define the quantities
(3.1)

Theorem 3.1. For a function ψ on T, the following statements are equivalent:

(a)  Mψ : w is bounded.

(b)  Mψ : 0w,0 is bounded.

(c)  θψ and ωψ are finite.

Furthermore, under the above conditions, one has

(3.2)

Proof. (a)(c) Assume Mψ is bounded from to w. The function fo = (1/2)χo and . Thus

(3.3)

Next, fix vT*. Then χv and ∥χv = 1; so

(3.4)
Taking the supremum over all vT, from (3.3) and (3.4) we see that ωψ is finite and
(3.5)

With vT*, we now define

(3.6)
Then fv, fv(o) = 0 and . By the boundedness of Mψ we obtain
(3.7)
Therefore,
(3.8)
Taking the supremum over all vT*, we obtain θψ < . From this and (3.5), we deduce the lower estimate
(3.9)

(c)(a) Assume θψ and ωψ are finite. Then, ψw, and by Lemma 1.1, for f with ∥f = 1 and vT*, we have

(3.10)
Thus, ψfw. Note that |f(o)| + ∥Df = 1 and
(3.11)
From this, we have
(3.12)
proving the boundedness of Mψ : w and the upper estimate
(3.13)

(b)(c) The proof is the same as for (a)(c); since for vT*, the functions χv and fv used there belong to 0.

(c)(b) Assume θψ and ωψ are finite, and let f0. Then, by Lemma 1.1, for vT*, we have

(3.14)
as |v | → . Thus, ψfw,0. The proof of the boundedness of Mψ is similar to that in (c)(a).

3.2. Isometries

In this section, we show there are no isometric multiplication operators Mψ from the space to w or from 0 to w,0.

Suppose Mψ : w is an isometry. Then . On the other hand,
(3.15)
Thus , which implies that ψ is a constant of modulus 1. Now observe that, for vT*, we have
(3.16)
which is a contradiction. Since χv0 for all vT, if Mψ : 0w,0 is an isometry, then the above argument yields again a contradiction. Thus, we proved the following result.

Theorem 3.2. There are no isometries Mψ from to w or from 0 to w,0.

3.3. Compactness and Essential Norm

We now characterize the compact multiplication operators, but first we give a useful compactness criterion for multiplication operators from to w or from 0 to w,0.

Lemma 3.3. A bounded multiplication operator Mψ from to w (0 to w,0) is compact if and only if for every bounded sequence {fn} in (0) converging to 0 pointwise, the sequence converges to 0 as n.

Proof. Suppose Mψ is compact from to w and {fn} is a bounded sequence in converging to 0 pointwise. Without loss of generality, we may assume for all n. Since Mψ is compact, the sequence {ψfn} has a subsequence that converges in the w-norm to some function fw.

By Lemma 1.3, for vT* we have

(3.17)
Thus, pointwise on T*. Furthermore, since , as k. Thus pointwise on T. Since by assumption, fn → 0 pointwise, it follows that f is identically 0, and thus . Since 0 is the only limit point in w of the sequence {ψfn}, we deduce that as n.

Conversely, suppose that every bounded sequence {fn} in that converges to 0 pointwise has the property that as n. Let {gn} be a sequence in such that for all n. Then |gn(o)| ≤ 1, and by part (a) of Lemma 1.1, for vT* we have |gn(v)|≤|v|. So {gn} is uniformly bounded on finite subsets of T. Thus there is a subsequence , which converges pointwise to some function g.

Fix ɛ > 0 and vT*. Then as well as for k sufficiently large. Therefore, for all k sufficiently large, we have

(3.18)
Thus g. The sequence is bounded in and converges to 0 pointwise. So as k. Thus in the w-norm. Therefore, Mψ is compact.

The proof for the case of Mψ : 0w,0 is similar.

Theorem 3.4. Let Mψ be a bounded multiplication operator from to w (or equivalently from 0 to w,0). Then the following are equivalent:

(a)  Mψ : w is compact.

(b)  Mψ : 0w,0 is compact.

(c)  lim |v|→ | v|2Dψ(v) = 0 and lim |v|→(|v | + 1) | ψ(v)| = 0.

Proof. (a)(c) Suppose Mψ : w is compact. We need to show that if {vn} is a sequence in T such that 2≤|vn| increasing unboundedly, then and lim n(|vn | + 1) | ψ(vn)| = 0. Let {vn} be such a sequence, and for n define . Clearly fn → 0 pointwise, and . Using Lemma 3.3, we see that

(3.19)

On the other hand, since fn(o) = 0 for all n, we have

(3.20)
Hence lim n(|vn | + 1) | ψ(vn)| = 0.

Next, for n, define

(3.21)
where ⌊x⌋ denotes the largest integer less than or equal to x. Then gn → 0 pointwise, and . Since , we have
(3.22)
By Lemma 3.3 we obtain lim n | vn|2Dψ(vn) ≤ lim nψgnw = 0.

(c)(a) Suppose lim |v|→|v|2Dψ(v) = 0 and lim |v|→(|v| + 1)|ψ(v)| = 0. Assume ψ is not identically zero, otherwise Mψ is trivially compact. By Lemma 3.3, to prove that Mψ is compact, it suffices to show that if {fn} is a bounded sequence in converging to 0 pointwise, then as n. Let {fn} be such a bounded sequence, , and fix ɛ > 0. There exists M such that (|v| + 1)|ψ(v)| < ɛ/2s and |v|2Dψ(v) < ɛ/2s for |v | ≥ M. For vT* and by Lemma 1.1, we have

(3.23)
Since fn → 0 uniformly on {vT:|v | ≤ M} as n, so does Dfn. So, on the set {vT:|v | ≤ M}, |v | D(ψfn)(v) → 0 as n. On the other hand, on {vT:|v | ≥ M}, we have
(3.24)
So |v | D(ψfn)(v) → 0 as n. Since fn → 0 pointwise, ψ(o)fn(o) → 0 as n. Thus as n. The compactness of Mψ follows at once from Lemma 3.3.

The proof of the equivalence of (b) and (c) is analogous.

For ψ a function on T, define
(3.25)

Theorem 3.5. Let Mψ be a bounded multiplication operator from to w. Then

(3.26)

Proof. Fix k, and for each n, consider the sets

(3.27)
Define the functions and . Then fn,k, gn,k0, , and fn,k and gn,k approach 0 pointwise as n. By Lemma 1.2, the sequences {fn,k} and {gn,k} approach 0 weakly in 0 as n. Let 𝒦0 be the set of compact operators from 0 to w,0, and note that every operator in 𝒦0 is completely continuous. Thus, if K𝒦0, then and , as n.

Therefore, if K𝒦0, then

(3.28)

Similarly,

(3.29)
Therefore, combining (3.28) and (3.29), we obtain
(3.30)
Letting k, we obtain .

Next, we wish to show that . The result is clearly true if (ψ) = 0. So assume there exists a sequence {vn} in T such that 2<|vn | → as n and

(3.31)
For n, define
(3.32)
Clearly, , and
(3.33)
The supremum of Dhn(v) is attained on the set {vT : |v| = |vn| − 1}. Thus . Define , and observe that gn0, ∥gn = 1, and gn → 0 pointwise on T. By Lemma 1.2, gn → 0 weakly in 0. Thus as n for any K𝒦0.

For each n, . Thus

(3.34)

We deduce that

(3.35)

Therefore,

(3.36)

We next derive an upper estimate on the essential norm.

Theorem 3.6. Let Mψ be a bounded multiplication operator from to w. Then

(3.37)

Proof. For each n, consider the operator Kn defined by

(3.38)
for f, where vn is the ancestor of v of length n. Then (Knf)(o) = f(o), and Knfw,0. Arguing as in the proof of Theorem 2.8, by the boundedness of Mψ, it follows that MψKn is a compact operator from to w.

Define the operator Jn = IKn, where I is the identity operator I on . Then,

(3.39)
Since (Jnf)(v) = 0 for |v | ≤ n, by Lemma 1.1, we obtain
(3.40)
From these two estimates, we arrive at
(3.41)
Since
(3.42)
from (3.41), taking the limit as n, we obtain
(3.43)
as desired.

4. Multiplication Operators from w or w,0 to L

In this section, we study the multiplication operators Mψ from the weighted Lipschitz space or the little weighted Lipschitz space into L. We begin by characterizing the bounded operators and determining their operator norm. In addition, we characterize the bounded operators that are bounded from below and show that there are no isometries among them. Finally, we characterize the compact multiplication operators and determine the essential norm.

4.1. Boundedness and Operator Norm

For a function ψ on T, define
(4.1)

Theorem 4.1. For a function ψ on T, the following statements are equivalent:

(a)  Mψ : wL is bounded.

(b)  Mψ : w,0L is bounded.

(c) is finite.

Furthermore, under the above conditions, one has ∥Mψ∥ = γψ.

Proof. The implication (a)(b) is obvious.

(b)(a) We begin by showing that, for each fw, the function ψf is bounded. Since Mψ is bounded on w,0, ψ = Mψ1 ∈ L. Thus, if f is constant, then ψfL. Fix fw, f nonconstant, v ∈ T, and set n = |v|. For wT, define

(4.2)
where wn is the ancestor of w of length n. Then fnw,0, and . Thus, ψfnL and
(4.3)
So |ψ(v)f(v)| = |ψ(v)fn(v)| ≤ ∥Mψ∥∥fw. Therefore, ψfL and
(4.4)
proving the boundedness of Mψ as an operator from w to L.

(a)(c) Assume Mψ : wL is bounded. Then ψ = Mψ1 ∈ L, and

(4.5)
For vT, define f(v) = log (1+|v|). Then f(o) = 0, and since for x ≥ 1 the function xxlog ((x + 1)/x) is increasing and has limit 1 as x, fw and ∥fw = 1. Thus
(4.6)
proving (c). Furthermore, from (4.5) and (4.6), we obtain
(4.7)

(c)(a) Assume . Let fw such that ∥fw = 1. Then |ψ(o)f(o)| ≤ |ψ(o)|, and by Lemma 1.3, for vT*, we have

(4.8)
Thus, ψfL and
(4.9)
proving the boundedness of Mψ as an operator from w to L. Taking the supremum over all functions fw such that ∥fw = 1, from (4.9) we obtain ∥Mψ∥ ≤ γψ. Therefore, from (4.7) we conclude that ∥Mψ∥ = γψ.

4.2. Boundedness from Below

Recall that an operator S from a Banach space 𝒳 to a Banach space 𝒴 is bounded below if there exists a constant C > 0 such that for all   xX
(4.10)

Theorem 4.2. A bounded multiplication operator Mψ from w or w,0 to L is bounded below if and only if

(4.11)

Proof. Assume Mψ is bounded below, and, arguing by contradiction, assume there exists vT such that ψ(v) = 0. Then Mψχv is identically 0. Since operators that are bounded below are necessarily injective [8], it follows that Mψ is not bounded below. Therefore, if Mψ is bounded below, then ψ is nonvanishing.

Next assume ψ is nonvanishing and inf vT |ψ(v)|/(|v | + 1) = 0. Then, there exists a sequence {vn} in T with 1≤|vn | → , such that |ψ(vn)|/(|vn | + 1)→0 as n. For n, define . Then ∥fnw = 1, but

(4.12)
Thus, Mψ is not bounded below.

Conversely, assume inf vT |ψ(v)|/(|v| + 1) = c > 0 and that Mψ is not bounded below. Then, for each n, there exists fnw such that ∥fnw = 1 and ∥ψfn < 1/n. Then, for each vT, we have

(4.13)
so that the sequence {gn} defined by gn(v) = (|v | + 1)fn(v) converges to 0 uniformly.

On the other hand, for vT*, we have

(4.14)
uniformly as n. Since |ψ(o)fn(o)| < 1/n, yet ∥fnw = 1, this yields a contradiction.

4.3. Isometries

In this section, we show there are no isometries among the multiplication operators from the spaces w or w,0 into L.

Suppose Mψ is an isometry from w or w,0 to L. Then, for vT the function fv = (1/(|v| + 1))χv is in w,0, , and
(4.15)
Thus, |ψ(v)| = |v| + 1. On the other hand, since Mψ is bounded, by Theorem 4.1, we have ; so ψ(v) → 0 as |v | → , which yields a contradiction. Thus, we proved the following result.

Theorem 4.3. There are no isometric multiplication operators Mψ from w or w,0 to L.

4.4. Compactness and Essential Norm

We begin by giving a useful compactness criterion for the bounded operators from w or w,0 into L.

Lemma 4.4. A bounded multiplication operator Mψ from w to L is compact if and only if for every bounded sequence {fn} in w converging to 0 pointwise, the sequence ∥ψfn approaches 0 as n.

Proof. Assume Mψ is compact on w, and let {fn} be a bounded sequence in w converging to 0 pointwise. By rescaling the sequence, if necessary, we may assume ∥fnw ≤ 1 for all n. By the compactness of Mψ, {fn} has a subsequence such that converges in the norm to some function fL. In particular, pointwise. Since by assumption, fn → 0 pointwise, it follows that f must be identically 0. Thus, the only limit point of sequence {ψfn} in L is 0. Hence ∥ψfn → 0.

Conversely, assume that for every bounded sequence {fn} in w converging to 0 pointwise, the sequence approaches 0 as n. Let {gn} be a sequence in w with ∥gnw ≤ 1. Fix wT, and, by replacing gn with gngn(w), assume gn(w) = 0 for all n. Then, for each vT, |gn(v)| = |gn(v) − gn(w)| ≤ d(v, w). Therefore, gn is uniformly bounded on finite subsets of T, and so some subsequence converges pointwise to some function g on T. Fix ɛ > 0 and vT*. Then, , , and for all k sufficiently large. Thus,

(4.16)
for k sufficiently large. Consequently, gw, we have
(4.17)
Since ɛ was arbitrary, it follows that ∥gw ≤ 1. Therefore, the sequence {fk} defined by is bounded in w and converges to 0 pointwise, hence, by the hypothesis, as n. We conclude that in L, proving the compactness of Mψ.

By an analogous argument, we obtain the corresponding compactness criterion for Mψ : w,0L.

Lemma 4.5. A bounded multiplication operator Mψ from w,0 to L is compact if and only if for every bounded sequence {fn} in w,0 converging to 0 pointwise, the sequence ∥ψfn approaches 0 as n.

Theorem 4.6. For a bounded operator Mψ from w to L (or equivalently from w,0 to L) the following statements are equivalent:

(a)  Mψ : wL is compact.

(b)  Mψ : w,0L is compact.

(c)  lim |v|→ log |v||ψ(v)| = 0.

Proof. (a)(b) is trivial.

(b)(c): Let {vn} be a sequence of vertices such that 1≤|vn | → as n. We need to show that

(4.18)
For n define
(4.19)
Then {fn} converges to 0 pointwise. Using the fact that (|v | − 1)(log  | v | − log (|v | − 1)) ≤ 1 for any choice of v in T* with |v | > 1, we have
(4.20)
for 2≤|v | ≤|vn|. Moreover, |v | Dfn(v) = 0 for |v | = 1 and for |v | >|vn|. Thus, fnw,0, and is bounded. By the compactness of Mψ as an operator from w,0 to L and by Lemma 4.5, we deduce
(4.21)
as n.

(c)(a) Assume {fn} is a sequence in w converging to 0 pointwise and such that a = sup n ∥fnw < . By Lemma 1.3, for all vT* and all n, we have

(4.22)
Fix ɛ > 0. There exists N such that N ≥ 3, and for |v | ≥ N, log  |v||ψ(v)| < ɛ/2a. Thus, for |v | ≥ N and for all n, |ψ(v)fn(v)| ≤ 2alog |v||ψ(v)| < ɛ. On the other hand, since fn → 0 pointwise, for each vertex v such that |v | < N and ψ(v) ≠ 0, we obtain |fn(v)| < ɛ/|ψ(v)| for all n sufficiently large. Hence |ψ(v)fn(v)| < ɛ for all vT and all n sufficiently large. Therefore, as n, which, by Lemma 4.4, proves the compactness of Mψ.

Next, we determine the essential norm of the bounded multiplication operators Mψ from w or w,0 to L.

Theorem 4.7. Let Mψ be a bounded multiplication operator from w or w,0 to L. Then

(4.23)

Proof. Define A(ψ) = lim n  sup |v|≥n log |v||ψ(v)|. If A(ψ) = 0, then by Theorem 4.6, Mψ is compact, hence its essential norm is 0. So assume A(ψ) > 0. We first show that . Let {vn} be a sequence in T such that 1≤|vn | → as n and

(4.24)
Fix p ∈ (0,1), and for each n, define
(4.25)
Then {fn,p} converges to 0 pointwise, fn,pw,0, fn,p(vn) = log  | vn|, and
(4.26)
By Lemma 1.4, {fn,p} converges to 0 weakly in w,0. Let K be a compact operator from w,0 (or equivalently, from w) to L. Since compact operators are completely continuous, it follows that as n. Thus,
(4.27)
Taking the infimum over all such compact operators K and passing to the limit as p approaches 0, we obtain
(4.28)

To prove the estimate , for each n and for fw, define

(4.29)
where vn is the ancestor of v of length n. In the proof of Theorem 2.8, it is was shown that Kn is a compact operator on w. Since Mψ : wL is bounded, it follows that MψKn is also compact as an operator from w to L.

Let vT, and let w be a vertex in the path from o to v of length k ≥ 1. Label the vertices from w to v by vj, j = k, …, |v|. Then for fw with ∥fw = 1, we have

(4.30)
Thus
(4.31)
We deduce
(4.32)

Taking the limit as n, we obtain .

5. Multiplication Operators from L to w or w,0

In this last section, we study the multiplication operators Mψ from L into the weighted Lipschitz space or the little weighted Lipschitz space. We first characterize the bounded operators and determine the operator norm. We also show there are no isometries among such operators. Finally, we characterize the compact multiplication operators and determine the essential norm.

5.1. Boundedness and Operator Norm

For a function ψ on T, define
(5.1)

Theorem 5.1. For a function ψ on T, the following statements are equivalent:

(a)  Mψ : Lw is bounded.

(b)  sup vT |v||ψ(v)| < .

Furthermore, under these conditions, one has

(5.2)

Proof. (a)(b) Assume Mψ : Lw is bounded. Fix vT*. Since χvL and , the function ψχvw, so

(5.3)
Thus, sup vT | vψ(v)| is finite.

(b)(a) Suppose sup vT|v||ψ(v)| < . Let fL such that ∥f = 1. Then

(5.4)
Thus, Mψ is bounded and ∥Mψ∥ ≤ ηψ.

We next show that ∥Mψ∥≥ηψ. The inequality is obvious if ψ is identically 0. For ψ not identically 0 and for vT, define

(5.5)
Then ∥f = 1, and for vT*, D(ψf)(v)   = |ψ(v)|+|ψ(v)|, so that
(5.6)
Thus, ∥Mψ∥≥ηψ, completing the proof.

In the next result, we characterize the bounded multiplication operators from L to w,0.

Theorem 5.2. For a function ψ on T, the following statements are equivalent:

(a)  Mψ : Lw,0 is bounded.

(b)  lim |v|→ |v||ψ(v)| = 0.

Furthermore, under these conditions, one has,

(5.7)

Proof. (a)(b) Assume Mψ : Lw,0 is bounded. Applying Mψ to the constant function 1, we obtain ψ = Mψ1 ∈ w,0. On the other hand, if 𝒪 = {vT : |v|  is  odd}, then ψχ𝒪w,0, so for vT*, we have

(5.8)
as |v | → , proving (b).

(b)(a) Suppose |v||ψ(v)| → 0 as |v | → . First observe that

(5.9)
as |v | → . Then for fL and vT*, we have
(5.10)
as |v | → . Thus, ψfw,0. The proof of the boundedness of Mψ and of the formula ∥Mψ∥ = ηψ is similar to the case when Mψ : Lw.

5.2. Isometries

As for all other multiplication operators in this paper, there are no isometries among the multiplication operators from L into w or w,0.

Assume Mψ is an isometry from L to w or w,0. Then, for vT the function fv = (1/(|v| + 1))χv is in w,0 with . In particular, it follows that |ψ(o)| = 1/2, and for vT*,
(5.11)
Thus, |ψ(v)| = 1/(|v | + 1). On the other hand, taking as a test function f the characteristic function of the set {vT:|v |    ≤ 1}, we obtain
(5.12)
which yields a contradiction. Therefore, we obtain the following result.

Theorem 5.3. There are no isometric multiplication operators Mψ from L to w or w,0.

5.3. Compactness and Essential Norm

The following two results are compactness criteria for multiplication operators from L into w or w,0 similar to those given in the previous sections.

Lemma 5.4. A bounded multiplication operator Mψ from L to w is compact if and only if for every bounded sequence {fn} in L converging to 0 pointwise, the sequence ∥ψfnw approaches 0 as n.

Proof. Assume Mψ is compact, and let {fn} be a bounded sequence in L converging to 0 pointwise. By rescaling the sequence, if necessary, we may assume ∥fn ≤ 1 for all n. By the compactness of Mψ, {fn} has a subsequence such that converges in the w-norm to some function fw. Since by Lemma 1.3, for vT*,

(5.13)
and , it follows that pointwise. Since by assumption, fn → 0 pointwise, the function f must be identically 0. Thus, the only limit point of the sequence {ψfn} in w is 0. Hence ∥ψfnw → 0 as n.

Conversely, suppose ∥ψfnw approaches 0 as n for every bounded sequence {fn} in L converging to 0 pointwise. Let {gn} be a sequence in L with . Then some subsequence converges to a bounded function g. Thus, the sequence converges to 0 uniformly, and is bounded. By the hypothesis, it follows that as k. Thus, in w. Therefore, Mψ is compact.

By an analogous argument, we obtain the corresponding result for Mψ : Lw,0.

Lemma 5.5. A bounded multiplication operator Mψ from L to w,0 is compact if and only if for every bounded sequence {fn} in L converging to 0 pointwise, the sequence approaches 0 as n.

Theorem 5.6. For a bounded operator Mψ from L to w, the following statements are equivalent:

(a)  Mψ is compact.

(b)  lim |v|→ |v||ψ(v)| = 0.

Proof. (a)(b) Assume Mψ is compact. Let {vn} be a sequence in T such that |vn | → as n. For n, let fn denote the characteristic function of the set {wT:|w | ≥|vn|}. Then and fn → 0 pointwise. By Lemma 5.4 and the compactness of Mψ, it follows that

(5.14)
as n.

(b)(a) Assume lim |v|→ |vψ(v)| = 0 and that ψ is not identically 0. In particular, ψ is bounded. Let {fn} be a sequence in L converging pointwise to 0 such that is bounded above by some positive constant C. Then corresponding to ɛ > 0, there exists N such that |v||ψ(v)| < ɛ/4C for all vertices v such that |v | ≥ N. Therefore, for |v | > N and n, we have

(5.15)
Furthermore, the sequence {fn} converges to 0 uniformly on the set {vT:|v | ≤ N} so that |fn(v)| < ɛ/4Nψ for all n sufficiently large. Hence |v | D(ψfn)(v) < ɛ for all vT* and all n sufficiently large. Consequently, as n. Using Lemma 5.4, we deduce that Mψ is compact.

Since the above proof is also valid when Mψ is a bounded operator from L to w,0, through the application of Lemma 5.5, from Theorems 5.2 and 5.6 we obtain the following result.

Corollary 5.7. For a function ψ on T, the following statements are equivalent:

(a)  Mψ : Lw is compact.

(b)  Mψ : Lw,0 is bounded.

(c)  Mψ : Lw,0 is compact.

(d) lim |v|→ |v||ψ(v)| = 0.

We now determine the essential norm of the bounded multiplication operators from L to w.

Theorem 5.8. Let Mψ : Lw be bounded. Then

(5.16)

Proof. Set B(ψ) = lim n  sup |v|≥n   | v | [|ψ(v)|+|ψ(v)|]. In the case B(ψ) = 0, then lim |v|→ |v | ψ(v) = 0, so by Theorem 5.6, Mψ is compact, and thus ∥Mψe = 0. So assume B(ψ) > 0. Then there exists a sequence {vn} in T such that 1≤|vn | → as n and

(5.17)
For each n let fn be the function on T defined by
(5.18)

Then , and {fn} converges to 0 pointwise. Thus, for any compact operator K : Lw, there exists a subsequence such that as k. Thus

(5.19)
Therefore, .

We now show that . For each n, define the operator Kn on L by

(5.20)
Then, for vT*, we have
(5.21)
Thus, Knfw with .

Assume {fk} is a sequence in L with . Then there exists a subsequence converging pointwise to some function fL. Thus,

(5.22)
So as j. Therefore, Kn is compact, and thus, since Mψ is bounded, MψKn is also compact.

For fL, we have

(5.23)
Therefore, we obtain
(5.24)
thus completing the proof.

Acknowledgment

The authors wish to thank the referee for the very helpful suggestions for the improvement of the paper.

      The full text of this article hosted at iucr.org is unavailable due to technical difficulties.