The Fréchet space of holomorphic functions on the unit disc
1 Introduction
The goal of this note is to develop all the machinery necessary to understand what it means to say that the set of holomorphic functions on the unit disc is a separable and reflexive Fréchet space that has the Heine-Borel property and is not normable.
2 Topological vector spaces
If is a topological space and , a local basis at is a set of open neighborhoods of such that if is an open neighborhood of then there is some that is contained in . We emphasize that to say that a topological vector space is normable is to say not just that there is a norm on the vector space , but moreover that the topology is induced by the norm.
A topological vector space over is a vector space over that is a topological space such that singletons are closed sets and such that vector addition and scalar multiplication are continuous. It is not true that a topological space in which singletons are closed need be Hausdorff, but one can prove that every topological vector space is a Hausdorff space.11 1 Walter Rudin, Functional Analysis, second ed., p. 11, Theorem 1.12. For any , we check that the map is a homeomorphism. Therefore, a subset of is open if and only if is open for all . It follows that if is a vector space and is a set of subsets of each of which contains , then there is at most one topology for such that is a topological vector space for which is a local basis at . In other words, the topology of a topological vector space is determined by specifying a local basis at . A topological vector space is said to be locally convex if there is a local basis at whose elements are convex sets.
If is a vector space and is a set of seminorms on , we say that is a separating family if implies that there is some with . (Thus, if is a seminorm on , the singleton is a separating family if and only if is a norm.) The following theorem presents a local basis at for a topology and shows that there is a topology for which the vector space is a locally convex space and for which this is a local basis at .22 2 Paul Garrett, Seminorms and locally convex spaces, http://www.math.umn.edu/~garrett/m/fun/notes_2012-13/07b_seminorms.pdf We call this topology the seminorm topology induced by .
Theorem 1 (Seminorm topology).
If is a vector space and is a separating family of seminorms on , then there is a topology on such that is a locally convex space and the collection of finite intersections of sets of the form
is a local basis at .
Proof.
We define to be those subsets of such that for all there is some satisfying . If is a subset of and , then there is some with , and there is some satisfying . We have
which tells us that . If and , then there are satisfying for . But the intersection of finitely many elements of is itself an element of , so , and
showing that . Therefore, is a topology.
Suppose that . For , let with ; there is such a seminorm because is a separating family. Then is an open set that contains and does not contain . Therefore is a closed set that contains and does not contain , and
is a closed set, showing that singletons are closed.
Let and . There are and , , such that . Let . If and , then there are and such that , and
so . This is true for each , , so . Hence
showing that vector addition is continuous at : for every basic open neighborhood of the image , there is an open neighborhood of whose image under vector addition is contained in .
Let , , and , say . Let , let be small enough so that for each , let , and let . If and , then
showing that . This is true for each , so , which shows that scalar multiplication is continuous at : for every basic open neighborhood of the image , there is an open neighborhood of whose image under scalar multiplication is contained in .
We have shown that with the topology is a topological vector space. To show that is a locally convex space it suffices to prove that each element of the local basis is convex. An intersection of convex sets is a convex set, so to prove that each element of is convex it suffices to prove that each is convex, and . If and , then
showing that and thus that is a convex set. Therefore, is a locally convex space. ∎
In the other direction, we will now explain how the topology of a locally convex space is induced by a separating family of seminorms. We say that a subset of a vector space is absorbing if implies that there is some such that . The Minkowski functional of an absorbing set is defined by
If is an open set containing and , then , and because scalar multiplication is continuous there is some such that . Thus an open set containing is absorbing. We say that a subset of a vector space is balanced if implies that . One proves that in a topological vector space, every convex open neighborhood of contains a balanced convex open neighhborhood of .33 3 Walter Rudin, Functional Analysis, second ed., p. 12, Theorem 1.14. It follows that a locally convex space has a local basis at whose elements are balanced convex open sets. The following lemma shows that the Minkowski functional of each member of this local basis is a seminorm.
Lemma 2.
If is a topological vector space and is a balanced convex open neighborhood of , then the Minkowski functional of is a seminorm on .
Proof.
Let and . If , then
Otherwise, write with and . Because is balanced and , we have
Therefore, if and , then .
Let . is absorbing, so let and . If then and . We have
and for , because is convex we have
where and , so
This is true for every , which means that . Therefore
showing that satisfies the triangle inequality and hence that is a seminorm on . ∎
We proved above that the Minkowski functional of a balanced convex open neighborhood of is a seminorm. The following lemma shows that the collection of Minkowski functionals corresponding to a balanced convex local basis at are a separating family.44 4 Walter Rudin, Functional Analysis, second ed., p. 27, Theorem 1.36.
Lemma 3.
If is a topological vector space and is a balanced convex open neighborhood of , then
If is a local basis at whose elements are balanced and convex, then
is a separating family of seminorms on .
Proof.
Let . If , then because and scalar multiplication is continuous, there is some and some open neighborhood of such that the image of under scalar multiplication is contained in . In particular, if and so . Thus we have
Therefore, if then . On the other hand, if and , then there is some such that . As is balanced, we have . Therefore, if then . This establishes that if then
If , then because singletons are closed, the set is open and contains , and thus there is some with . Hence , which implies by the first claim that . In particular, , proving the second claim. ∎
If is a locally convex space then there is a local basis at , call it , whose elements are balanced and convex, and we have established that is a separating family of seminorms on . Therefore by Theorem 1, with the seminorm topology induced by is a locally convex space. The following theorem states that the seminorm topology is equal to the original topology of the space.55 5 Paul Garrett, Seminorms and locally convex spaces, http://www.math.umn.edu/~garrett/m/fun/notes_2012-13/07b_seminorms.pdf
Theorem 4.
If is a locally convex space, then there is a separating family of seminorms on such that is equal to the seminorm topology.
Proof.
Let be a local basis at whose elements are balanced and convex and let . If , then , which is an open neighborhood of in the seminorm topology induced by , and this implies that the seminorm topology is at least as fine as .
If and , then
and , and it follows that is at least as fine as the seminorm topology. Therefore is equal to the seminorm topology induced by . ∎
We have shown that if is a vector space and is a separating family of seminorms on , then with the seminorm topology induced by is a locally convex space. Furthermore, we have shown that if is a locally convex space then there is a separating family of seminorms on such that the topology of is equal to the seminorm topology induced by . In other words, the topology of any locally convex space is the seminorm topology induced by some separating family of seminorms on the space.
A subset of a topological vector space is said to be bounded if for every open neighborhood of there is some such that implies that .
Lemma 5.
If is a locally convex space with the seminorm topology induced by a separating family of seminorms on , then a subset of is bounded if and only if each is a bounded function on .
Proof.
Suppose that is bounded and . The set is an open neighborhood of , so there is some such that . Hence if then , so is a bounded function on .
Suppose that for each there is some such that implies that . If is an open neighborhood of , then there are and such that
Let . For ,
i.e.,
But if and then
hence is in the above intersection and thus is in . Therefore , showing that is bounded. ∎
We now prove that if the topology of a locally convex space is induced by a countable separating family of seminorms then the topology is metrizable.
Theorem 6.
If is a locally convex space with the seminorm topology induced by a countable separating family of seminorms and is a summable nonincreasing sequence of positive numbers, then
is a translation invariant metric on , is equal to the metric topology for , and with this metric the open balls centered at are balanced.
Proof.
For any we have
because the sequence is summable. It is apparent that .
If is any seminorm on , then
so
Also, it is straightforward to check that the function defined by satisfies . Define . If , then
Hence, for ,
showing that satisfies the triangle inequality.
If , then
As each term is nonnegative, each term must be equal to . As each is positive, this implies that each is equal to . But is a separating family so if then there is some with , and this shows that , i.e. . Therefore is a metric on .
If , then : the metric is translation invariant.
If and , then
Thus, if and then , so the open ball
is balanced.
has a local basis at whose elements are finite intersections of sets of the form . Suppose that , let be large enough so that , and let be large enough so that . If , then
This shows that
and this entails that is at least as fine as the metric topology induced by .
Suppose that and . If , then of course for each we have
and hence if then
and hence if then
Therefore,
It follows from this that the metric topology induced by is at least as fine as . ∎
If a locally convex space is metrizable with a complete metric, then it is called a Fréchet space.
We now prove conditions under which a topological vector space is normable.
Theorem 7.
A topological vector space is normable if and only if there is a convex bounded open neighborhood of the origin.
Proof.
Suppose that is a convex bounded open neighborhood of . contains a balanced convex open neighborhood of ,66 6 Walter Rudin, Functional Analysis, second ed., p. 12, Theorem 1.14. and because is bounded so is . We define , where is the Minkowski functional of . If , then because is an open neighborhood of and is bounded, there is some such that . Hence , i.e., . As is balanced, by Lemma 3 we get . is a seminorm, so , showing that if then , and hence that is a norm on . Also, we check that
Because is bounded, for any open neighborhood of there is some such that , hence
This implies that the norm topology for is at least as fine as . And is an open set because scalar multiplication is continuous, so is at least as fine as the norm topology for . Therefore that is normable with the norm .
In the other direction, if is the norm topology for some norm on , then
is indeed a convex open neighborhood of the origin. Suppose that is an open neighborhood of . There is some such that
and thus such that , and hence is bounded, showing that there exists a convex bounded open neighborhood of the origin. ∎
A topological vector space is called locally bounded if there is a bounded open neighborhood of the origin. A topological vector space is said to have the Heine-Borel property if every closed and bounded subset of it is compact.
Theorem 8.
If is a topological vector space that is locally bounded and has the Heine-Borel property, then has finite dimension.
Proof.
Let be a bounded neighborhood of . It is a fact that the closure of a bounded set is itself bounded,77 7 Walter Rudin, Functional Analysis, second ed., p. 11, Theorem 1.13(f). and therefore is compact. For any , the set is a compact neighborhood of , hence is locally compact. But a locally compact topological vector space is finite dimensional,88 8 Walter Rudin, Functional Analysis, second ed., p. 17, Theorem 1.22. so is finite dimensional. ∎
3 Continuous functions on the unit disc
Let , the open unit disc. Let be the set of continuous functions . is a complex vector space. If is a compact subset of , define
It is straightforward to check that is a seminorm on . If is nonzero then there is some with , and hence , so the set of all is a separating family of seminorms on . Thus, with the seminorm topology induced by the set of all is a locally convex space.
Define , . If is a compact subset of , then there is some with , so , and hence
It follows that the seminorm topology induced by is at least as fine as the seminorm topology induced by , thus the topologies are equal. Because the topology of is induced by the countable family , by Theorem 6 it is metrizable: for any summable nonincreasing sequence of positive real numbers , the topology is induced by the metric
(1) |
Suppose that is a Cauchy sequence. For , the fact that is a Cauchy sequence in implies that as . is a Banach space with the norm , and hence there is some satisfying as . We define to be , for ; this makes sense because the restriction of to is if . is continuous at each point in because for each point in there is some containing an open neighborhood of the point, and is continuous. Hence . Therefore with the metric (1) is a complete metric space, which means that it is a Fréchet space.
Theorem 9.
The topology of is not induced by a norm.
Proof.
Because the topology of is the seminorm topology induced by the separating family of seminorms , by Lemma 5 a subset of is bounded if and only if each is a bounded function on , i.e., for each there is some such that implies .
Suppose by contradiction that there is a bounded convex open neighborhood of the origin. Because for any , there is some and some such that
being bounded implies that is bounded. Let
and let be a partition of unity subordinate to this open cover of . For any constant , the restriction of to is and hence belongs to . But , so is not a bounded function on , contradicting that is bounded. Therefore, there is no bounded convex open neighborhood of . By Theorem 7, this tells us that is not normable. ∎
For each , the set is a Banach space with norm . If and , let be the restriction of to . For , the function is a continuous linear map , and if then . Thus the Banach spaces and the maps are a projective system in the category of locally convex spaces, and it is a fact that any projective system in this category has a projective limit that is unique up to unique isomorphism.
Theorem 10.
.
Proof.
Define by taking to be the restriction of to . Each is continuous and linear. Certainly, if then . Suppose the is a locally convex space, that are continuous linear maps, and that if then
(2) |
If and , then by (2) we have . For , eventually , and define to be for any such that . For each there is some such that is in the interior of , and the restriction of to is equal to , hence is continuous at . Therefore , so .
Suppose that and . If , then there is some with , and because is linear,
Therefore is linear.
Suppose that is a net with limit . For to converge to means that for each we have . But
and because is continuous. Therefore, for each we have , so is continuous. ∎
We proved in the above theorem that the Fréchet space is the projective limit of the Banach spaces . It is a fact that the projective limit of any projective system of Banach spaces is a Fréchet space.99 9 J. L .Taylor, Notes on locally convex topological vector spaces, http://www.math.utah.edu/~taylor/LCS.pdf, p. 8, Proposition 2.6, and cf. Paul Garret, Functions on circles: Fourier series, I, http://www.math.umn.edu/~garrett/m/fun/notes_2012-13/04_blevi_sobolev.pdf, p. 37, §13.
A topological space is said to be separable if it has a countable subset that is dense.
Theorem 11.
is separable.
Proof.
One proves using the Stone-Weierstrass theorem that the Banach space is separable. The product of at most continuum many separable Hausdorff spaces each with at least two points is itself separable with the product topology.1010 10 Stephen Willard, General Topology, p. 109, Theorem 16.4. Therefore, is separable. Because each is a metric space, this countable product is metrizable, and any subset of a separable metric space is itself separable with the subspace topology. The projective limit of a projective system of topological vector spaces is a closed subspace of the product of the spaces; thus, using merely that the projective limit is a subset of the product and has the subspace topology inherited from the direct product, we get that is separable. ∎
4 Holomorphic functions on the unit disc
Let be the set of holomorphic functions . is a linear subspace of . Let have the subspace topology inherited from . One proves that this topology is equal to the seminorm topology induced by . Any subset of a separable metric space with the subspace topology is separable. By Theorem 11 the Fréchet space is separable, and thus is separable too.
We now prove that is a closed subspace of .1111 11 Paul Garrett, Holomorphic vector-valued functions, http://www.math.umn.edu/~garrett/m/fun/notes_2012-13/08b_vv_holo.pdf A closed linear subspace of a Fréchet space is itself a Fréchet space, hence this theorem shows that is a Fréchet space.
Theorem 12.
is a closed subset of .
Proof.
Suppose that is a net and that . We shall show that . (In fact it suffices to prove this for a sequence of elements in because we have shown that is metrizable, but that will not simplify this argument.) To show this we have to prove that if then has a limit as , . Let be a counterclockwise oriented circle contained in with center , say of radius . For each the function is holomorphic on , and so Cauchy’s integral formula gives
Therefore
As is a compact subset of this gives us
The right-hand side tends to 0, while the left-hand side does not depend on . Hence
(3) |
Applying (3), we have for ,
hence
thus
For we have , and so by the dominated convergence theorem we get
Thus, for every , the function is complex differentiable at . Hence , and therefore is a closed subset of . ∎
We remind ourselves that a topological vector space is said to have the Heine-Borel property if every closed and bounded subset of it is compact. Lemma 5 tells us that a subset of is bounded if and only if each seminorm is a bounded function on . The following theorem states that has the Heine-Borel property.1212 12 Henri Cartan, Elementary Theory of Analytic Functions of One or Several Complex Variables, pp. 162–167, chapter V, §4. An equivalent statement is called Montel’s theorem.
Theorem 13 (Heine-Borel property).
The Fréchet space has the Heine-Borel property.
That has the Heine-Borel property is a useful tool, and lets us prove that the topology of is not induced by a norm.
Theorem 14.
is not normable.
Proof.
If were normable then by Theorem 7 there would be a convex bounded open neighborhood of the origin. This would imply that is locally bounded (has a bounded open neighborhood of the origin). But has the Heine-Borel property, and a topological vector space that is locally bounded and has the Heine-Borel property is finite dimensional by Theorem 8. It is straightforward to check that is not finite dimensional, and hence is not normable. ∎
For , let . First, if then one proves that . Then, the following theorem states that is a morphism in the category of locally convex spaces.1313 13 Henri Cartan, Elementary Theory of Analytic Functions of One or Several Complex Variables, p. 143, chapter V, §1.
Theorem 15.
Differentiation is a continuous linear map.
If is a compact subset of and , let be the restriction of to , and let be the closure in of the set . Each element of is holomorphic on the interior of . is a Banach space with the norm , and hence is a Banach space with the same norm, because it is indeed a linear subspace. If and , let . The are continuous and linear, and if then . Thus the Banach spaces and the continuous linear maps are a projective system in the category of locally convex spaces, and this projective system has a projective limit . The following theorem states that this projective limit is equal to the Fréchet space .1414 14 J. L .Taylor, Notes on locally convex topological vector spaces, http://www.math.utah.edu/~taylor/LCS.pdf, p. 8
Theorem 16.
.
5 Dual spaces
The dual of a topological vector space is the set of continuous linear maps . If is a bounded subset of and , then is a bounded subset of (the image of a bounded set under a continuous linear map is a bounded set). Hence
The function is a seminorm on , and if then there is some with , hence . The strong dual topology on is the seminorm topology induced by the separating family
(To add to our vocabulary: the set of all bounded subsets of a topological vector space is called the bornology of the space. Similar to how one can define a topology as a collection of sets satisfying certain properties, one can also define a bornology on a set without first having the structure of a topological vector space.) We denote by the dual space with the strong dual topology. is a locally convex space. If is a normed space, one can prove1515 15 K. Yosida, Functional Analysis, sixth ed., p. 111, Theorem 1. that is normable with the operator norm
We say that a topological vector space is reflexive if ; since the strong dual of a topological vector space is locally convex, for a topological vector space to be reflexive it is necessary that it be locally convex.
Let be a locally convex space. The Hahn-Banach separation theorem1616 16 Walter Rudin, Functional Analysis, second ed., p. 59, Theorem 3.4. yields that separates : if then there is some with . If , then is a seminorm on and is therefore a separating family of seminorms on . We call the seminorm topology induced by this separating family the weak topology on , and with the weak topology is a locally convex space. The original topology on is at least as fine as the weak topology on : any set that is open using the weak topology is open using the original topology.
The following lemma shows that a Fréchet space with the Heine-Borel property is reflexive, and therefore that is reflexive.
Lemma 17.
If a Fréchet space has the Heine-Borel property, then it is reflexive.
Proof.
A subset of a locally convex space is called a barrel if it is closed, convex, balanced, and absorbing. A locally convex space is said to be barreled if each barrel is a neighborhood of . It is a fact that every Fréchet space is barreled.1717 17 K. Yosida, Functional Analysis, sixth ed., p. 138, Corollary 1. A locally convex space is reflexive if and only if it is barreled and if every set that is closed, convex, balanced, and bounded is weakly compact.1818 18 K. Yosida, Functional Analysis, sixth ed., p. 140, Theorem 2. Therefore, for a Fréchet space with the Heine-Borel property to be reflexive it is necessary and sufficient that every set that is compact, convex, and balanced be weakly compact. But if a subset of a locally convex space is compact then it is weakly compact, because the original topology is at least as fine as the weak topology and hence any cover of a set by elements of the weak topology is also a cover of the set by elements of the original topology. Therefore, any Fréchet space with the Heine-Borel property is reflexive. ∎
Morphisms in the category of locally convex spaces are continuous linear maps. If and are locally convex spaces and is a morphism, the dual of is the morphism
defined by
One verifies that is in fact a morphism. If the spaces and the morphisms , , are a projective system in the category of locally convex spaces, then the dual spaces and the morphisms , , are a direct system in this category. It is a fact that the dual of a projective limit of Banach spaces is isomorphic to the direct limit of the duals of the Banach spaces.1919 19 Paul Garrett, Functions on circles: Fourier series, I, http://www.math.umn.edu/~garrett/m/fun/notes_2012-13/04_blevi_sobolev.pdf, p. 15, Theorem 5.1.1. Thus, as is the projective limit of the Banach spaces , its dual space is isomorphic to the direct limit of the duals of these Banach spaces:
Cooper2020 20 J. B. Cooper, Functional analysis– spaces of holomorphic functions and their duality, http://www.dynamics-approx.jku.at/lena/Cooper/holloc.pdf, p. 11, §5. shows that is isomorphic to the space of germs of functions on the complement of in the extended complex plane that vanish at infinity. Let be those sequences satisfying
By Hadamard’s formula for the radius of convergence of a power series, these are precisely the sequences of coefficients of power series with radius of convergence , and is a complex vector space. The map
is linear and has the linear inverse
so and are linearly isomorphic. For , define
Each is a norm, yet we do not give the norm topology. Rather, we give the seminorm topology induced by the family , and with this topology is a locally convex space. One proves that the above two linear maps are continuous, and hence that is isomorphic as a locally convex space to . Then, one proves that the dual space of are those sequences such that
and corresponds to