nLab Haar integral

Redirected from "Haar measure".

Context

Measure theory

Group Theory

Contents

Idea

For GG a locally compact Hausdorff topological group, with C c(G)C_c(G) denoting its algebra of compactly supported continuous functions from GG to \mathbb{R}, a Haar integral G\int_G is a continuous linear map C c(G)C_c(G) \rightarrow \mathbb{R} to the real numbers which is invariant under the obvious group action of GG on C c(G)C_c(G). It turns out that there exists a unique such integral, up to a scalar multiple. The Riesz representation theorem then allows one to conclude the existence of a unique Haar measure, which is a GG-invariant Borel measure on GG.

The archetypal example of Haar measure is the Lebesgue measure on the (additive group underlying) Cartesian space n\mathbb{R}^n.

Definition

Let GG be a locally compact Hausdorff topological group. Let C c(G)C_c(G) denote the vector space of continuous real-valued functionals with compact support on GG. This is a locally convex topological vector space where the locally convex structure is specified by the family of seminorms

ρ K(f)=sup xK|f(x)|, \rho_K(f) \;=\; \sup_{x \in K} {|f(x)|} \,,

where KK ranges over compact subsets of GG. Recall that a Radon measure on GG may be described as a continuous linear functional

G:C c(G). \int_G \,\colon\, C_c(G) \to \mathbb{R} \,.

Such a Radon measure defines a measure μ\mu on the σ \sigma -algebra of Borel sets in the usual sense of measure theory, where

μ(B)=sup{ Gf:supp(f)=KB,ρ K(f)=1}. \mu(B) \;=\; sup \Big\{ \textstyle{\int_G} f \,\colon\, supp(f) = K \subseteq B,\, \rho_K(f) = 1 \Big\} \,.

In this context, a (left) Haar integral on GG is a nonzero such linear functional G\int_G such that

Gf0whenf0 \int_G f \;\geq\; 0 \;\;\text{when}\; f \geq 0
Gf g= Gf \int_G f^g \;=\; \int_G f

for each fC c(G)f \in C_c(G) and each gGg \in G, where f g:Gf^g \colon G \rightarrow \mathbb{R} sends xx to f(gx)f(g x).

Correspondingly, a (left) Haar measure on GG is a nonzero Radon measure μ\mu such that

μ(gB)=μ(B) \mu(g B) \;=\; \mu(B)

for all gGg \in G and all Borel sets BB.

Properties

Existence and Uniqueness

Any locally compact Hausdorff topological group GG admits a Haar integral (and therefore Haar measure) that is unique up to scalar multiple. This result was first proven by André Weil. Proofs may be found in Alfsen 1963 (consructive) and Rubinstein-Salzedo 2004.

A particular argument for the case of a compact Hausdorff topological group was provided by Bader 2020, which is what we follow now:

Theorem

Let GG be a compact Hausdorff topological group. Write C 0(G)C^0(G) for the Banach space of continuous \mathbb{R}-valued functions on GG. There is a unique GG-invariant \mathbb{R}-linear map of Banach-spaces :C 0(G)\int \colon C^0(G) \rightarrow \mathbb{R} such that 0f 20 \leq \int f^2 for all fC 0(G)f \in C^0(G).

Proof

Consider the set SS of subsets CC of C 0(G) *C^0(G)^* which satisfy

(1) (GG-invariance) for all functionals ϕ\phi in CC, the continuous functional ϕ g\phi^g sending ff to ϕ(f g 1)\phi(f^{g^{-1}}) is contained in CC.

(2) (compactness) The subtopology on CC is compact.

(3) (convexity) for elements ϕ 1,,ϕ n\phi_1, \ldots, \phi_n in CC and a 1,...,a n[0,)a_1, ..., a_n \in [0,\infty) such that i=1 na i=1\sum_{i = 1}^n a_i = 1, i=1 na iϕ iC\sum_{i = 1}^n a_i \phi_i \in C.

(4) (nonemptiness) there is an element of C 0(G) *C^0(G)^* contained in CC.

Claim 1: SS is nonempty.

Recall that the Hahn-Banach theorem theorem for a Banach space with squared Minkowski seminorm sup iIϕ iϕ i¯\text{sup}_{i \in I} \phi_i \overline{\phi_i} for continuous functionals {ϕ i} iI\{ \phi_i \}_{i \in I} provides the existence of a continuous functional ϕ:C 0(G)\phi : C^0(G) \rightarrow \mathbb{R}. {ϕ}\{ \phi \} satisfies (4) and (2) but possibly not (3) or (1).

The \infty-norm topology on C 0(G)C^0(G) given by the squared Minkowski seminorm sup gGev gev g¯\text{sup}_{g \in G} ev_g \overline{ev_g} is equivalently described as the compact open topology on continuous functions from GG to \mathbb{R}.

For any fC 0(G)f \in C^0(G), {f}C 0(G)\{ f \} \subset C^0(G) is closed because it is a single point in a Hausdorff space. Write μ g 1:GG\mu_{g^{-1}} : G \rightarrow G for the continuous function sending hh to hg 1hg^{-1}. For a continuous function f:Gf : G \rightarrow \mathbb{R}, write f g:Gf^g : G \rightarrow \mathbb{R} for fμ g 1f \circ \mu_{g^{-1}}.

The composition

ev(1×(μ(1×Inv))):[G,]×G×G \text{ev} \circ \Big(1 \times \big(\mu \circ (1 \times \text{Inv}) \big) \Big) \;\colon\; [G,\mathbb{R}] \times G \times G \longrightarrow \mathbb{R}

is continuous, from which we get that

[G,]×G[G,] [G,\mathbb{R}] \times G \longrightarrow [G,\mathbb{R}]

is continuous, using the calculus of mates for exponentiable objects in topological spaces, along with the fact that a compact space is exponentiable. [G,]×G[G,][G,\mathbb{R}] \times G \rightarrow [G,\mathbb{R}] sending (f,g)(f,g) to the continuous function f gf^g sending hh to f(hg 1)=fμ g 1(h)f(hg^{-1}) = f \circ \mu_{g^{-1}} (h) is continuous. It follows that image of {f}×G\{ f \} \times G in [G,][G,\mathbb{R}] is compact. This is the GG-orbit {f g:gG}\{ f^g : g \in G \}. {f g:gG}\{ f^g : g \in G \} satisfies (1) (2) and (4), but possibly not (3).

The Krein-Smulian theorem, applied to GfG \cdot f which is compact in the Banach space C 0(G) *C^0(G)^*, gives that the convex hull Gf¯\overline{G \cdot f} of GfG \cdot f is compact in the weak topology on C 0(G) *C^0(G)^* (not to be confused with the weak *{}^* topology on C 0(G) *C^0(G)^*). We can conclude that Gf¯\overline{G \cdot f} is GG-invariant, compact (relative to the weak topology), convex, and nonempty. Therefore Gf¯S\overline{G \cdot f} \in S.

Claim 2: given a chain {C n} n\{ C_n \}_{n \in \mathbb{N}} of elements C nSC_n \in S such that C n+1C nC_{n+1} \subset C_n, nC nS\cap_{n \in \mathbb{N}} C_n \in S.

(1) (GG-invariance) if ϕ nC n\phi \in \cap_{n \in \mathbb{N}} C_n, then ϕC n\phi \in C_n for each nn \in \mathbb{N}, so ϕ gC n\phi^g \in C_n for each nn \in \mathbb{N}, so ϕ g nC n\phi^g \in \cap_{n \in \mathbb{N}} C_n.

(2) (compactness) nC n\cap_{n \in \mathbb{N}} C_n is closed so

(3) (convexity) nC n\cap_{n \in \mathbb{N}} C_n is closed so

(4) (nonemptiness) C nC_n is a decreasing chain of compact subsets of a Hausdorff topological space, from which we can conclude that it is nonempty.

Claim 3: Zorn's Lemma for the case of the partial order defined on SS by which C 1C 2C_1 \leq C_2 if and only if C 2C 1C_2 \subseteq C_1 is satisfied because of claim 2 and claim 3. This implies that we may find a minimal element of SS in this collection, where SS is ordered where ABA \leq B when ABA \subset B.

Choose such an element CC.

Claim 4: CC is a singleton.

The Krein-Milman theorem, applied to the compact, convex subset CC of C 0(G) *C^0(G)^* (again as a locally convex topological vector space under the weak topology), gives that CC is the convex hull of its extreme points.

The continuous group action of GG on CC sends the subspace Inext(C)\text{Inext}(C) of inextremal points to itself, as the average i=1 na if i\sum_{i = 1}^n a_i f_i of f if_i with a i0a_i \neq 0, a i(0,)a_i \in (0,\infty), and i=1 na i\sum_{i = 1}^n a_i remains an average with respect to the same weights a i0a_i \neq 0, a i(0,)a_i \in (0,\infty), with i=1 na i\sum_{i = 1}^n a_i:

(a 1f 1+a 2f 2) g=a 1f 1 g+a 2f 2 g (a_1 f_1 + a_2 f_2)^g = a_1 f_1^g + a_2 f_2^g

Since GG acts by homeomorphisms on CC, GG must send the subspace of extremal points Ext(C)\text{Ext}(C) of CC to itself. Using again that GG acts by homeomorphisms, we can conclude that ϕ\phi is extremal if and only if ϕ g\phi^g is extremal.

For any fInext(C)f \in \text{Inext}(C), one can construct an object in SS using the facts above (the compactness of any GG-orbit and the resulting compactness of its convex hull C˜\tilde{C} within C 0(X) *C^0(X)^*), and from that the inclusions {f}Inext(C)C\{ f \} \subset \text{Inext}(C) \subsetneq C imply inclusions of the GG-orbits G{f}Inext(C)CG \cdot \{ f \} \subset \text{Inext}(C) \subsetneq C, which imply inclusions G{f}¯Inext(C)C\overline{G \cdot \{ f \}} \subset \text{Inext}(C) \subseteq C, from which we can conclude that CC is not minimal, a contradiction!

Therefore 0=card(Inext(C))0 = \text{card}(\text{Inext}(C)). This implies 1=card(Ext(C))1 = \text{card}(\text{Ext}(C)), i.e. CC is a singleton, and its unique element a GG-invariant functional extending ff.

Extensive and Intensive Properties

In Lawvere’s thinking about extensive and intensive quantities,

  • C(G)C(G) is a space of intensive quantities on GG.

  • [C(G),] Ban[C(G), \mathbb{R}]_{\text{Ban}} is the space of extensive quantities on XX, where Ban\text{Ban} is the category of Banach spaces with bounded maps as maps. The Haar integral is the unique GG-invariant element of this space of norm 11.

  • the integration map is the canonical evaluation pairing

    G:C(G)×[C(G),] Ban. \int_G \;\colon\; C(G) \times [C(G), \mathbb{R}]_{\text{Ban}} \longrightarrow \mathbb{R} \,.

If we suggestively write Gfdϕ\int_G f d \phi for G(f,ϕ)=ϕ(f)\int_G (f, \phi) = \phi(f), then Gdϕ\int_G - d \phi becomes a way of writing ϕ\phi. In particular, choosing ϕ\phi to be the Haar measure, we can write ϕ\phi as Gdϕ\int_G - d \phi.

Left and Right Haar Measures that Differ

The left and the right Haar measure may or may not coincide, groups for which they coincide are called unimodular. Consider the matrix subgroup

G:={(y x 0 1)|x,y,y>0} G := \left\{ \left.\, \begin{pmatrix} y & x \\ 0 & 1 \end{pmatrix}\,\right|\, x, y \in \mathbb{R}, y \gt 0 \right\}

The left and right invariant measures are, respectively,

μ L=y 2dxdy,μ R=y 1dxdy \mu_L = y^{-2} \,\mathrm{d}x \,\mathrm{d}y,\quad \mu_R = y^{-1} \,\mathrm{d}x \,\mathrm{d}y

and so G is not unimodular.

Abelian groups are obviously unimodular; so are compact groups and discrete groups.

References

Textbook account:

Proofs:

  • E. M. Alfsen: A Simplified Constructive Proof of the Existence and Uniqueness of Haar Measure, Mathematica Scandinavica 12 (1963) [doi:10.7146/math.scand.a-10675]

  • Simon Rubinstein-Salzedo: On the existence and uniqueness of invariant measures on locally compact groups (2004) [pdf]

  • Uri Bader, MO answer (2020) [MO:a/351405]

See also:

Last revised on June 26, 2025 at 22:53:13. See the history of this page for a list of all contributions to it.