§ 2. Support of a measure
1. Restriction of a measure to an open set. Definition of a measure by means of local data
Let X be a locally compact space, Y an open set in X. The subspace Y of X is locally compact, and every continuous function with values in a topological vector space E, defined on Y and with compact support, may be extended by continuity to all of X, by giving it the value 0 on ; one can therefore identify in this way the space with the linear subspace of formed by the continuous functions with compact support contained in Y. If is a measure on X, it is clear that the restriction of to is a measure on Y, which is called the restriction of to the open subspace Y, or the measure induced on Y by , and is denoted . The restrictions to Y of , and are, respectively, , and , by virtue of §1, Nos. 5 and 6. If is real then the restrictions of and to Y are, respectively, and , by virtue of formula (8) of §1, No. 5.
One sees immediately that if Y and Z are two open sets in X such that , and if and are the restrictions of to Y and Z, then is also the restriction of to the open subspace Z of the locally compact space Y.
In Ch. IV, §5, No. 7 we shall generalize this definition to the case that Y is a locally compact subspace of X.
Note that a measure on Y is not necessarily the restriction of some measure on X (cf. Ch. V, §7, No. 2, Prop. 3).
For example, let Y be the open interval ]0,1[ of ; the mapping
is a measure on , because every function in is zero on a neighborhood of 0 in . However, this measure cannot be extended to a measure on because, in the contrary case, its restriction to the set of functions such that would be bounded; but this is false.
However, we have the following proposition:
Proposition 1
Let* be an open covering of and suppose given, on each subspace , a measure , in such a way that for every pair , the restrictions of and to are identical. Under these conditions, there exists one and only one measure on whose restriction to is equal to for every index .
Let us first show that every function may be written in the form of a finite sum where, for each of the functions , there exists an index such that . If , there exists a finite number of indices () such that the form a covering of ; let () be continuous mappings of into such that the support of is compact and is contained in for , and such that on (§ 1, No. 2, Lemma 1); the functions meet the requirements. This shows in the first place that if there exists a measure meeting the requirements then it is unique, because for every finite sum , where , necessarily . Moreover, we will have shown the existence of a linear form on whose restriction to each subspace is , provided we demonstrate the following property: if and are two finite sequences of functions in such that for , for and
whence
It remains to see that is a measure on ; now, every point of admits a compact neighborhood contained in one of the ; the conclusion therefore follows at once from the definition of and from Prop. 6 of §1, No. 3.
Corollary (Principle of localization). — Let and be two measures on , and let be a family of open sets of such that, for every , the restrictions to of and are equal; then the restrictions of and to are equal.
2. Support of a measure
Let be a measure on a locally compact space , and let be the set of open sets such that the restriction of to is zero; it follows at once from the principle of localization (No. 1, Cor. of Prop. 1) that if is the union of the sets , then itself belongs to and is therefore the largest of the sets of .
Definition 1
If is a measure on a locally compact space , one defines the support of , denoted , to be the closed set complementary to the largest of the open sets in on which the restriction of is zero.
To say that a point does not belong to the support of means that there exists an open neighborhood of such that the restriction of to is zero; to say that belongs to the support of therefore means that for every neighborhood of , there exists a function , whose support is contained in , such that .
Example
— 1) For a measure on to be zero, it is necessary and sufficient that its support be empty. 2) The support of Lebesgue measure on is the entire line ; for, it is nonempty and is invariant under every translation.
3) On the interval of consider a countable dense subset, arranged into a sequence , and let be the measure defined by placing the mass at the point for every (§ 1, No. 3, Example I). The support of is all of ; for, let be any point of , a neighborhood of , and a continuous real-valued function on , equal to 1 at the point , whose support is contained in (§ 1, No. 2, Lemma 1); the set of such that is open in , therefore contains a point , consequently .
Proposition 2
The support of a measure is identical to the support of the measure ; if is real, its support is the union of the supports of the measures and .
For, if the restriction of to an open set is zero, then so is that of (resp. of and when is real), and conversely.
Note that the supports of and can be nonempty and identical (cf. Ch. V, § 5, Exer. 5).
Proposition 3
If and are two measures on a locally compact space such that , then .
For, if the restriction of to an open set is zero, then so is that of .
Proposition 4
The support of the sum of two measures is contained in the union of their supports.
For, if the restrictions of two measures to an open set are zero, then the same is true of the restriction of their sum.
If and are two positive measures, then the support of is equal to the union of the supports of and ; for, if is a point of the union and is any neighborhood of , then there exists a continuous function with support contained in and such that one of the numbers , is ; a fortiori, .
Proposition 5
The support of the restriction of a measure to an open set is the trace on of the support of .
The proposition is obvious from the definitions.
Proposition 6
The set of measures on a locally compact space , whose support is contained in a closed set , is a vaguely closed linear subspace of .
For, it is the intersection of the vaguely closed hyperplanes with equation , where runs over the set of functions in whose support does not intersect .
Suppose is not compact: given a filter on the space of measures on , we shall say that the support of a measure recedes indefinitely along if, for every compact subset of , there exists a set such that for every measure .
Proposition 7
— *If is a filter on such that the support of recedes indefinitely along , then converges vaguely to 0 with respect to .
Let be any function in and let be its support. By hypothesis, there exists a set such that for every measure , ; it follows that for all , which proves the proposition.
3. Characterization of the support of a measure
By definition, if the support of a function does not intersect the support of a measure , then ; but the following more precise result is true:
Proposition 8
— *Let be a measure on a locally compact space . For every function that is zero on , .
Set , . Given a number , let be the set of such that ; is an open set containing by hypothesis, therefore is a neighborhood of the compact set . It follows that there exists a continuous mapping of into , equal to 1 on and with support contained in (§ 1, No. 2, Lemma 1). Since the support of does not intersect , . On the other hand, on , and on , therefore on , by the choice of . Observe finally that there exists a number such that for every function whose support is contained in ; since the support of is contained in , and consequently ; since is arbitrary, .
Corollary 1
— *If two functions in are equal on , then .
Corollary 2
— *Let be a positive measure on ; if is such that on , then .
For, on , therefore by Corollary 1.
Corollary 3
— *Let be a bounded measure on ; if is such that on , then .
For, , and if is a continuous mapping of into , equal to 1 on and with compact support, then on , therefore
by Corollary 2; the conclusion then follows from formula (13) of § 1, No. 6.
Proposition 9
— Let be a positive measure on ; if is a function in such that , then is zero on .
Let be a point of such that ; let us show that does not belong to . For, there exist a compact neighborhood of and a number such that on . If is any continuous function with support contained in , let us show that ; indeed, if one sets then , whence .
Proposition 10
— Let be a measure on a locally compact space ; for every function , the support of the measure is the closure of the set of points such that .
Set ; let be a point not belonging to ; there exists an open neighborhood of such that at every point of , is zero; if has support contained in , then is zero on , therefore (Prop. 8) ; in other words, the restriction of to is zero.
Conversely, assuming that the restriction of to an open neighborhood of a point is zero, let us show that there does not exist a point of at which is . Indeed, if there were such a point , there would exist a compact neighborhood of , contained in , at every point of which ; but then every function with support contained in could be written , where has support contained in ; it would then follow that , contrary to the hypothesis .
Note that is contained in the intersection of the support of and the support of , but it is not necessarily equal to this intersection. For example, if , is the Dirac measure at the point 0, and , then even though the intersection of the supports of and reduces to the point 0, thus is nonempty.
Corollary
— In order that the measure be zero, it is necessary and sufficient that be zero on the support of .
Proposition 11
— Every measure with compact support is bounded.
For, is also a measure with compact support, thus we can restrict attention to the case that ; if is a continuous mapping of into , with compact support and equal to 1 on , then for every function one has on , therefore (Cor. 2 of Prop. 8) , which proves the proposition (§ 1, No. 8).
4. Point measures. Measures with finite support
Proposition 12
— Let () be distinct points in a locally compact space . Every measure on whose support is contained in the set of the is a linear combination of the measures ().
For, such a measure is zero for every function satisfying the relations (No. 3, Prop. 8); since these relations may be written , is a linear combination of the (A, II, §7, No. 5, Cor. 1 of Th. 7).
In particular, every measure whose support is either empty or reduced to a single point is of the form , where is a complex number; such a measure is said to be a point measure; thus every measure whose support is finite is a sum of point measures.
Theorem 1
— Every measure on a locally compact space is in the vague closure of the vector space of measures whose support is finite and contained in .
It suffices to prove that is orthogonal to the subspace of orthogonal to (TVS, II, §6, No. 3, Cor. 2 of Th. 1), that is, that the relations , where runs over the support of , imply ; but this is just Prop. 8 of No. 3.
Corollary 1
— Every bounded measure on is in the vague closure of the convex set of measures whose support is finite and is contained in that of , and whose norm is . Moreover, if tends vaguely to while remaining in , then tends to .
To prove the first assertion, it suffices to establish that the measure belongs to the polar set of the polar set of in (TVS, II, §6, No. 3, Th. 1 and §8, No. 4); this means that for , the relations for all imply that ; but this is a consequence of Cor. 3 of Prop. 8 of No. 3.
To prove the second assertion, we note that
since the function is lower semi-continuous for the vague topology (§1, No. 9, Cor. 4 of Prop. 15), and the conclusion follows from the fact that for by definition.
Corollary 2
— Every bounded measure on is in the vague closure of the set of measures whose support is finite and contained in that of and whose norm is equal to .
We can suppose that . Let be an open neighborhood of 0 for the vague topology; for every such that , there exists, by virtue of Cor. 1, a measure whose support is finite and contained in and for which and . Setting , we have and ; for sufficiently small we therefore have , whence the conclusion.
Corollary 3
Every bounded positive measure on is in the vague closure of the convex set of positive measures whose support is finite and contained in that of and whose norm is equal to .
The same reasoning as in Cor. 2 shows that we can limit ourselves to proving that is in the vague closure of the convex set formed by the positive measures with finite support contained in and with norm . Again, it suffices to establish that belongs to the polar set of , the polar set of in the space (TVS, II, §6, No. 3, Th. 1); but this means that for the relations for all imply , which is a consequence of No. 3, Cor. 2 of Prop. 8.
Corollary 4
In the space , the set of point measures is total for the topology of strictly compact convergence* (§1, No. 10).
On the cone , the topology of strictly compact convergence is identical to the vague topology (§1, No. 10, Prop. 18), and every measure on may be written , where the () are positive measures; the conclusion therefore follows from Th. 1.
Proposition 13
Let be a measure on a locally compact space . For a point to belong to , it is necessary and sufficient that the point measure be in the vague closure of the set of measures , where runs over the set of continuous functions with compact support such that .
The condition is obviously sufficient by virtue of Prop. 6 of No. 2. To see that it is necessary, suppose ; consider a finite number of functions () in , and an arbitrary number ; we are to prove that there exists a function such that and such that
multiplying by a scalar, we can suppose that . This being so, setting we can write, for and for every function ,
Since has its support in , we may identify it with its restriction to ; the hypothesis then implies that there exists a function , with support contained in , such that and such that for . The definition of moreover shows that for all ; since and we therefore have and so, setting ,
This proves the proposition, since .
Corollary
Let be a measure on . For a measure on to be in the vague closure of the set of measures , where runs over , it is necessary and sufficient that .
For, by No. 3, Prop. 10; therefore the support of every vague limit of measures of the form is also contained in (No. 2, Prop. 6). Conversely, if then is the vague limit of measures with finite support contained in (Th. 1), hence is in the vague closure of the set of measures by Prop. 13.
5. Discrete measures
Proposition 14
For a measure on a locally compact space to be a discrete measure (§ 1, No. 3, Example I), it is necessary and sufficient that be a discrete closed subspace of .
Let be a discrete measure on , defined by the masses placed at the points of a discrete closed subspace of ; let us show that . For every and every neighborhood of , there exists a function with support contained in , equal to 1 at the point and to 0 at the other points of , whence . On the other hand if then there exists a neighborhood of not intersecting ; for every function with support contained in , we therefore have , which proves that .
Conversely, let be a measure such that is a discrete closed subspace of . For every , there exists an open neighborhood of that contains no point of distinct from ; the restriction of to is therefore a point measure with support (No. 2, Prop. 5), hence (No. 4, Prop. 12) is of the form , where . Setting at the points of , and denoting by the measure defined by the masses , the principle of localization shows that .
We thus see that on a discrete space , every measure is discrete.
Exercises
See the exercises for § 2.