Let G be a locally compact group. It will be understood, once and for all, that the measures μ1,…,μn on G are said to be convolvable if they are so for the mapping
(x1,x2,…,xn)↦x1x2⋯xn;
and it is by means of this mapping that the convolution product ∗iμi will always be taken. If s∈G,t∈G, then
(1)
εs∗εt=εst.
If s∈G and μ∈M(G), then
(2)
εs∗μ=γ(s)μ
(3)
μ∗εs=δ(s−1)μ
by §1, No. 1, Example 3. If G is abelian, to say that μ1 and μ2 are convolvable is equivalent to saying that μ2 and μ1 are convolvable, and one then has μ1∗μ2=μ2∗μ1. When G is not abelian, it can happen that μ1 and μ2 are convolvable, without μ2 and μ1 being so (Exer. 12).
Proposition 1
— Let G be a locally compact group, λ,μ,ν measures =0 on G.
(i) If λ,μ,ν are convolvable, then so are λ and μ, ∣λ∣∗∣μ∣ and ν, μ and ν, λ and ∣μ∣∗∣ν∣, and one has
λ∗μ∗ν=(λ∗μ)∗ν=λ∗(μ∗ν).
(ii) If λ and μ are convolvable, as well as ∣λ∣∗∣μ∣ and ν, then λ,μ,ν are convolvable. Similarly if μ and ν are convolvable, as well as λ and ∣μ∣∗∣ν∣.
This follows from Prop. 1 of §1, No. 2.
There can exist measures λ,μ,ν on G such that the convolution products λ∗μ, (λ∗μ)∗ν, μ∗ν, λ∗(μ∗ν) are all defined, and yet (λ∗μ)∗ν=λ∗(μ∗ν) (cf. Exer. 4).
Let ρ be a lower semi-continuous finite function >0 on G such that ρ(st)≤ρ(s)ρ(t) for all s,t in G. We denote by Mρ(G) the vector space of measures λ on G such that ρ is λ-integrable, and by ∥λ∥ρ (or simply ∥λ∥) the norm ∫Gρ(s)d∣λ∣(s) on this space. When ρ=1, one recovers the set M1(G) of bounded measures on G.
Proposition 2
(i) Any two elements of Mρ(G) are convolvable. (ii) For convolution, and for the norm ∥λ∥, Mρ(G) is a complete normed algebra, admitting εe as unity element. (iii) C′(G)is a subalgebra of Mρ(G).
Let λ,μ be in Mρ(G), and let us show that λ and μ are convolvable. Let f∈K+(G). Since ρ is >0 and lower semi-continuous, there exists a constant k>0 such that f≤kρ. Then
∫∗f(st)d∣λ∣(s)d∣μ∣(t)≤k∫∗ρ(st)d∣λ∣(s)d∣μ∣(t)
≤k∫∗ρ(s)ρ(t)d∣λ∣(s)d∣μ∣(t)
=k(∫∗ρ(s)d∣λ∣(s))(∫∗ρ(t)d∣μ∣(t))
(Ch. V, §8, No. 3, Cor. 1 of Prop. 8). Therefore (s,t)↦f(st) is (λ⊗μ)-integrable, so that λ and μ are convolvable. On the other hand, using Ch. V (§1, Prop. 4, §6, Prop. 2, §8, Cor. 1 of Prop. 8) and the fact that (s,t)↦ρ(s)ρ(t) is lower semi-continuous in G×G, one has
One sees that λ∗μ∈Mρ(G) and that ∥λ∗μ∥≤∥λ∥⋅∥μ∥. In view of Prop. 1, Mρ(G) is an algebra. The mapping λ↦ρ⋅λ is an isometric linear mapping θ of Mρ(G) into M1(G); if μ∈M1(G) then 1/ρ, which is locally bounded and upper semi-continuous, is locally μ-integrable, and ρ is (1/ρ)⋅μ-integrable, thus (1/ρ)⋅μ∈Mρ(G); this proves that θ is surjective; therefore Mρ(G) is a complete normed algebra. Finally, it is clear that εe is a unity element for Mρ(G) and that C′(G) is a subalgebra of Mρ(G) (§1, No. 4, Cor. of Prop. 5).
If ρ=1, Prop. 2, (i) and (ii) also follow from §1, Prop. 2.
Proposition 3
Let μ1,…,μn be measures on G. If all of the μi, except at most one, have compact support, then the μi are convolvable.
For, let Si be the support of μi, and suppose that Si is compact for i=i0. Let K be a compact subset of G. The set of (x1,…,xn)∈∏iSi such that x1x2⋯xn∈K is compact, because the conditions xi∈Si for all i, x1x2⋯xn∈K imply
xi0∈Si0−1−1⋯S1−1KSn−1⋯Si0+1−1.
Therefore the μi are convolvable (§1, No. 4, Prop. 4).
Proposition 4
The mapping (λ,μ)↦λ∗μ (resp. (λ,μ)↦μ∗λ), where λ∈C′(G), μ∈M(G), defines on M(G) the structure of a left (resp. right) module over the algebra C′(G).
This follows from Props. 1 and 3.
Proposition 5
Let λ be a left (resp. right) Haar measure on G, and μ∈M1(G). Then μ and λ (resp. λ and μ) are convolvable, and μ∗λ=μ(1)λ (resp. λ∗μ=μ(1)λ).
We can suppose that μ⩾0. Let f∈K+(G). When λ is a left Haar measure,
∫∗dμ(x)∫∗f(xy)dλ(y)=∫∗dμ(x)∫f(y)dλ(y)=λ(f)∥μ∥,
therefore the function (x,y)↦f(xy) is (μ⊗λ)-integrable, and its integral for μ⊗λ is λ(f)∥μ∥. One argues similarly when λ is a right Haar measure.
Proposition 6
Let μ and ν be two convolvable measures on G. Let χ be a continuous representation of G in C∗. Then χ⋅μ and χ⋅ν are convolvable and (χ⋅μ)∗(χ⋅ν)=χ⋅(μ∗ν).
Let f∈K(G). Then fχ∈K(G), therefore the function
(x,y)↦f(xy)χ(xy)=f(xy)χ(x)χ(y)
on G×G is integrable for μ⊗ν; therefore the function (x,y)↦f(xy) is integrable for (χ⋅μ)⊗(χ⋅ν); therefore χ⋅μ and χ⋅ν are convolvable. Moreover,
Let G and G' be two locally compact groups, u a continuous representation of G in G', and μ1,…,μn measures on G, all =0. Then the following assertions are equivalent: (i) u is μi-proper for all i, and the measures u(∣μi∣) are convolvable; (ii) the μi are convolvable and u is proper for ∗i(∣μi∣). When these conditions are satisfied,
∗iu(μi)=u(∗iμi).
This follows from §1, No. 2, Cor. of Prop. 1.
Corollary
Let G be a locally compact group, μ1,…,μn measures on G. For the sequence (μi)1≤i≤n to be convolvable, it is necessary and sufficient that the sequence (μˇn−i)0≤i≤n−1 be so, in which case
(μ1∗⋯∗μn)∨=μˇn∗⋯∗μˇ1.
This follows from Prop. 7 on considering the isomorphism x↦x−1 of G onto the opposite group.
2. The case of a group operating on a space
Let X be a locally compact space on which a locally compact group G operates on the left continuously by
(s,x)↦s⋅x.
If μ1,…,μn are measures on G and ν is a measure on X, these will be said to be convolvable if they are so for the mapping (s1,…,sn,x)↦s1⋯snx of Gn×X into X, and their convolution product is to be understood in the sense of this mapping.
If s∈G and x∈X, then
(4)
εs∗εx=εsx.
If s∈G and μ∈M(X), then
(5)
εs∗μ=γ(s)μ
by §1, No. 1, Example 3.
Proposition 8
— Let μ be a measure on G, ν a measure on X. (i) If μ has compact support, then μ and ν are convolvable. (ii) If ν has compact support, and if G operates properly in X, then μ and ν are convolvable. This follows from Prop. 4 of §1, No. 4.
Proposition 9
— For convolution, M1(X) is a left module over M1(G), while M(X) and C′(X) are left modules over C′(G). This follows from Prop. 8, and from §1, Props. 1, 3 and the Cor. of Prop. 5.
Proposition 10
— Let μ be a measure on G, ν a measure on X, μ and ν being convolvable. Suppose in addition that there exists a positive measure β on X such that γ(s)ν has base β for every s∈G. Then μ∗ν has base β. Let K be a β-negligible compact subset of X. Then K is γ(s)∣ν∣-negligible for every s∈G. Now,
∣μ∣∗∣ν∣=∫G(εs∗∣ν∣)d∣μ∣(s)
(§1, No. 5, Prop. 7), and the mapping s↦εs∗∣ν∣ is vaguely continuous (§2, Prop. 6). Therefore K is ∣μ∣∗∣ν∣-negligible by Ch. V, §3, No. 3, Th. 1. Therefore ∣μ∣∗∣ν∣ has base β (Ch. V, §5, No. 5, Th. 2).
3. Convolution and linear representations
Proposition 11
— Let G be a locally compact group, E a quasi-complete locally convex space, U a continuous representation of G in E. (i) If λ∈C′(G), μ∈C′(G), then U(λ∗μ)=U(λ)U(μ). (ii) Suppose that E is a Banach space, and let ρ(s)=∥U(s)∥ for s∈G. If λ∈Mρ(G), μ∈Mρ(G), then U(λ∗μ)=U(λ)U(μ). Let λ,μ be in C′(G). For any x∈E one has, by applying notably Props. 1 and 4 of Ch. VI, §1, No. 1,
U(λ∗μ)x=∫GU(s)xd(λ∗μ)(s)
=∫G×GU(st)xdλ(s)dμ(t)=∫G×GU(s)U(t)xdλ(s)dμ(t)
=∫GU(λ)U(t)xdμ(t)=U(λ)∫GU(t)xdμ(t)=U(λ)U(μ)x,
whence (i). An analogous argument may be applied in case (ii).
With G still a locally compact group, let us assume that G operates continuously on the left in a locally compact space X. This defines (§2, No. 4) a continuous linear representation γ of G in M(X) (equipped with the topology of compact convergence in H(X)).
Proposition 12
If λ∈C′(G) and μ∈M(X), then
γ(λ)μ=λ∗μ.
By Prop. 7 of §1, No. 5,
λ∗μ=∫G(εs∗μ)dλ(s).
Now, εs∗μ=γ(s)μ (No. 2, formula (5)) and
∫G(γ(s)μ)dλ(s)=γ(λ)μ
by the definition of γ(λ).
Corollary
The mapping (λ,μ)↦λ∗μ of C′(G)×M(X) into M(X) is hypocontinuous relative to the equicontinuous subsets of C′(G) and the compact subsets of M(X) (C′(G) and M(X) being equipped with the topology of compact convergence in C(G) and H(X), respectively).
For, M(X), equipped with the topology of compact convergence in H(X), is quasi-complete. Therefore the mapping (λ,μ)↦γ(λ)μ of C′(G)×M(X) into M(X) is hypocontinuous relative to the equicontinuous subsets of C′(G) and the compact subsets of M(X) (§2, No. 6). It then suffices to apply Prop. 12.
Remark
— 1) Let λ0∈C′(G). The mapping μ↦λ0∗μ of M(X) into M(X) is vaguely continuous. For, let f∈H(X). One has
⟨λ0∗μ,f⟩=∫f(sx)dλ0(s)dμ(x)=⟨μ,g⟩,
where g(x)=∫f(sx)dλ0(s). Now, g is continuous (Ch. VII, §1, No. 1, Lemma 1). On the other hand, let S be the support of λ0 and K that of f. The conditions sx∈K and s∈S imply x∈S−1K; therefore the support of g is contained in S−1K, so that g∈H(X). Then ⟨λ0∗μ,f⟩=⟨μ,g⟩ is a vaguely continuous function of μ, which proves our assertion.
2) Let μ0∈M(X). The mapping λ↦λ∗μ0 of C′(G) into M(X) is continuous for the topologies σ(C′(G),C(G)) and σ(M(X),H(X)). For, let f∈K(X). Setting h(s)=∫f(sx)dμ0(x), we have ⟨f,λ∗μ0⟩=⟨h,λ⟩, and h∈C(G) (Ch. VII, §1, No. 1, Lemma 1).
Proposition 13
The mapping* (s,μ)↦γ(s)μofG×M+(X)intoM+(X)is continuous when the setM+(X)of positive measures onXis equipped with the vague topology.
Since γ(s)μ=γ(ss0−1)γ(s0)μ, it follows from Remark 1 that it suffices to prove the continuity of the mapping under consideration at a point of the form (e,μ0) with μ0∈M+(X). Given a function f∈K(X) and a number ε>0, it is thus a matter of showing that there exist a neighborhood U of e in G and a neighborhood W of μ0 in M+(X) such that the relations s∈U,μ∈W imply
∫f(sx)dμ(x)−∫f(x)dμ0(x)≤ε.
Let V be a compact neighborhood of the support K of f in X, and let φ∈K+(X) be such that φ(x)=1 on V; there exists a neighborhood W0 of μ0 in M+(X) such that a=supμ∈W0μ(V) is finite: it suffices to take for W0 the set of μ∈M+(X) such that ∣⟨φ,μ−μ0⟩∣≤1. Since the mapping (s,x)↦sx is continuous, there is, on the other hand, a compact neighborhood U0 of e in G such that sK⊂V for all s∈U0; the function (s,x)↦f(sx) is then uniformly continuous in U0×V and so there is a neighborhood U⊂U0 of e such that ∣f(sx)−f(x)∣≤ε/2a for all s∈U and x∈V. For s∈U and μ∈W0, we therefore have
∫f(sx)dμ(x)−∫f(x)dμ(x)≤ε/2;
if W⊂W0 is the neighborhood of μ0 in M+(X) formed by the measures μ∈W0 such that ∫f(x)dμ(x)−∫f(x)dμ0(x)≤ε/2, U and W meet the requirements.