Professional Documents
Culture Documents
Lecture Notes
Originally Created for the Class of Winter Semester 2016/2017 at LMU Munich
Includes Subsequent Corrections and Revisions†
Contents
1 Measure Theory 4
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2 Measure Spaces and Generators . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 Extension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.3.1 Semirings, Rings, Algebras . . . . . . . . . . . . . . . . . . . . . . 11
1.3.2 Contents, Premeasures . . . . . . . . . . . . . . . . . . . . . . . . 14
1.3.3 Lebesgue Premeasure on Rn . . . . . . . . . . . . . . . . . . . . . 22
1.3.4 Outer Measures, Carathéodory Extension Theorem . . . . . . . . 27
1.3.5 Dynkin Systems, Uniqueness of Extensions . . . . . . . . . . . . . 31
1.3.6 Completion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
1.4 Inverse Image, Trace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
1.5 Lebesgue Measure on Rn . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
1.5.1 Intervals and Countable Sets . . . . . . . . . . . . . . . . . . . . . 41
∗
E-Mail: philip@math.lmu.de
†
Resources used in the preparation of this text include [Els07, For17, Kön04].
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
CONTENTS 2
1.5.2 Approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
1.5.3 Cantor Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
1.6 Measurable Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
1.6.1 Definition, Composition . . . . . . . . . . . . . . . . . . . . . . . 46
1.6.2 Pushforward Measure . . . . . . . . . . . . . . . . . . . . . . . . . 48
1.6.3 Invariance of Lebesgue Measure under Euclidean Isometries . . . 49
1.6.4 Nonmeasurable Sets . . . . . . . . . . . . . . . . . . . . . . . . . 54
1.6.5 R-Valued Measurable Maps . . . . . . . . . . . . . . . . . . . . . 56
1.6.6 Products, Projections, and Borel Sets . . . . . . . . . . . . . . . . 65
2 Integration 69
2.1 Integration of R-Valued Measurable Maps . . . . . . . . . . . . . . . . . 69
2.1.1 Simple Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
2.1.2 Nonnegative Functions . . . . . . . . . . . . . . . . . . . . . . . . 72
2.1.3 Integrable Functions . . . . . . . . . . . . . . . . . . . . . . . . . 75
2.2 Null Sets, Properties Holding Almost Everywhere . . . . . . . . . . . . . 81
2.3 Convergence Theorems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
2.3.1 Fatou’s Lemma and Dominated Convergence . . . . . . . . . . . . 82
2.3.2 Parameter-Dependent Integrals: Continuity, Differentiation . . . . 84
2.4 Riemann versus Lebesgue . . . . . . . . . . . . . . . . . . . . . . . . . . 87
2.5 Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
2.5.1 Product Measure . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
2.5.2 Theorems of Tonelli and Fubini . . . . . . . . . . . . . . . . . . . 101
2.6 Lp -Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
2.6.1 Definition, Norm, Completeness . . . . . . . . . . . . . . . . . . . 108
2.6.2 Integration with Respect to Pushforward Measures . . . . . . . . 116
2.6.3 Dense Subsets, Separability . . . . . . . . . . . . . . . . . . . . . 117
2.6.4 Convolution and Fourier Transform . . . . . . . . . . . . . . . . . 123
2.7 Change of Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
CONTENTS 3
References 200
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 4
1 Measure Theory
1.1 Motivation
In Analysis I, we have already measured the size of intervals I := [a, b] ⊆ R by defining
|I| := |a−b|. Moreover, we defined the Riemann integral of Riemann-integrable functions
f : I −→ R on such intervals. We R also mentioned that, for a Riemann-integrable f :
+
I −→ R0 , the Riemann integral I f can be interpreted as the size of the area between
the graph of f and the horizontal axis. Major goals of this class include measuring the
size of subsets of Rn and integrating functions over subsets of Rn .
Q
While for n-dimensional intervals I = nj=1 Ij ⊆ Rn Q with one-dimensional intervals
Ij ⊆ R, an obvious choice to assign a volume is |I| := nj=1 |Ij |, for general subsets of
Rn , it is far from obvious what a reasonable volume should be. For certain sets that
have a simple relation with intervals, this relation can be used to assign a volume (e.g.
triangles can be decomposed into halves of rectangles, which leads to the formula for its
area that is taught in high school). Another common sense rule is that the volume v of
the union of N disjoint sets with known volume vj should be the sum of the volumes,
P
v= N j=1 vj . And it makes sense to extend this rule to countable disjoint unions: For
example, if i1 1 i
∀ Ij := j , j−1 ,
j∈N 2 2
then the Ij are clearly disjoint,
∞
[
I := [0, 1] = {0} ∪ Ij ,
j=1
and, due to the formula for the sum of the geometric series,
∞
X ∞
X
1 −j
|I| = 1 = 1 −1=0+ 2 = |[0, 0]| + |Ij |.
1− 2 j=1 j=1
For example, such a rule then allows to compute the area of a circle by decomposing it
into countably many triangles (however, the details are a bit cumbersome and we will
not pursue this here).
Another common sense property that a volume function should satisfy is translation
invariance, i.e. v(A + x) = v(A) should hold for each A ⊆ Rn and each x ∈ Rn , where
A + x := {a + x : a ∈ A}.
Definition 1.1. Let n ∈ N. We call a function v : P(Rn ) −→ [0, ∞] an ideal n-
dimensional volume if, and only if, the following conditions (i) – (iii) are satisfied:
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 5
Unfortunately, using the axiom of choice, we will be able to show that no ideal volume
exists (see Sec. 1.6.4 and, in particular, Th. 1.74(b)). However, we will see that there
exist volume functions (notably so-called Lebesgue-Borel and Lebesgue measure) that
satisfy Def. 1.1(a)–(c), except that they are not defined on all of P(Rn ) (Lebesgue-Borel
measure is defined on a strictly smaller set than Lebesgue measure and is the restriction
of Lebesgue measure to this set, cf. Ex. 1.47).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 6
(a) ∅ ∈ A.
(b) If A ∈ A, then Ac = X \ A ∈ A.
S
(c) If (An )n∈N is a sequence of sets in A, then n∈N An ∈ A.
If A is a σ-algebra on X, then the pair (X, A) is called a measurable space. The sets
A ∈ A are called A-measurable (or merely measurable if A is understood).
—
Measures will be defined on σ-algebras (see Def. 1.10 below). Note that there are
similarities between the definition of a σ-algebra and the definition of a topology: For
example, both topologies and σ-algebras are always subsets of the power set of some
given set of interest; they also both contain at least the empty set and the whole space.
Of course, in general, a topology is not a σ-algebra and a σ-algebra is not a topology.
Still, we will notice repeatedly, that similar concepts and techniques are useful for both
measure spaces and topological spaces (there is a branch of mathematics called Category
Theory that systematically investigates such aspects – one finds that the category of
topological spaces has a lot in common with the category of measurable spaces).
(b) A is closed under countable (both finite and infinite) intersections: Let I be a
nonempty countable index set and Aj ∈ A, j ∈ I. Then
\
Aj ∈ A.
j∈I
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 7
(a) Clearly, {∅, X} and P(X) are σ-algebras on X, where {∅, X} sometimes called the
trivial σ-algebra on X.
(b) A := {A ⊆ X : A countable or Ac countable} constitutes a σ-algebra S on X
c
(clearly, ∅ ∈ A; if A ∈ A, then A ∈ A; if An ∈ A, n ∈ N,Tthen A := n∈N An ∈ A:
If at least one Acn0 , n0 ∈ N, is countable, then Ac = n∈N Acn ⊆ Acn0 is count-
able; if each Acn is uncountable, then each An must be countable, implying A to be
countable).
is a again a σ-algebra on X.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 8
(a) Both {∅} and {X} are generators of the trivial σ-algebra on X. The set E := {{x} :
x ∈ X} generates the σ-algebra defined in Ex. 1.4(b) above.
(b) Let T be a topology on X. Then B := σ(T ), i.e. the σ-algebra generated by the
open sets of X, is called the Borel σ-algebra on (X, T ) (or on X if the topology T
is understood). The elements of B are called Borel sets.
(c) Let n ∈ N. It is often useful to know that the Borel σ-algebra B n on Rn (with
respect to the norm topology on Rn ) is also generated by each of the following sets
(in each case, it is a simple consequence of the following Lem. 1.9):
Cn := {A ⊆ Rn : A closed},
Kn := {A ⊆ Rn : A compact},
Icn := {[a, b] : a, b ∈ Rn , a ≤ b} ∪ {∅},
n
Ic,Q := {[a, b] : a, b ∈ Qn , a ≤ b} ∪ {∅},
Ion := {]a, b[: a, b ∈ Rn , a ≤ b},
n
Io,Q := {]a, b[: a, b ∈ Qn , a ≤ b},
I n := {]a, b] : a, b ∈ Rn , a ≤ b},
IQn := {]a, b] : a, b ∈ Qn , a ≤ b},
S
IΣn := { N n
k=1 Ik : N ∈ N, I1 , . . . , IN ∈ I disjoint},
S
n
IΣ,Q := { N n
k=1 Ik : N ∈ N, I1 , . . . , IN ∈ IQ disjoint}.
Lemma 1.9. Let (X, T ) be a topological space and let B := σ(T ) be the corresponding
set of Borel sets. Moreover, let C be the set of closed subsets of X, and let K be the set
of compact subsets of X. Then the following holds true:
(a) B = σ(C).
(b) If (X, T ) is Hausdorff (i.e. Sa T2 space) and there exists a sequence (Kn )n∈N of
compact sets such that X = ∞ n=1 Kn , then B = σ(K).
(c) If E ⊆ B is such that every open set O ∈ T is a countable union of sets from E,
then B = σ(E).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 9
i.e. A is a countable union of sets from K, showing C ⊆ σ(K) and B ⊆ σ(K), proving
(b).
(c): The hypothesis of (c) implies T ⊆ σ(E) and B ⊆ σ(E). As E ⊆ B is assumed, this
already proves (c).
Keeping in mind our considerations of the previous section, we are now ready to define
what we mean by a measure:
(i) µ(∅) = 0.
(ii) µ is countably additive (also called σ-additive), i.e., if (Ai )i∈N is a sequence in
A consisting of (pairwise) disjoint sets, then
∞
! ∞
[ X
µ Ai = µ(Ai ). (1.1)
i=1 i=1
If µ is a measure on (X, A), then the triple (X, A, µ) is called a measure space. In
the context of measure spaces, the sets A ∈ A are sometimes called µ-measurable
instead of A-measurable.
(b) Let (X, A, µ) be a measure space. The measure µ is called finite or bounded if, and
only if, µ(X) < ∞ (for µ(X) = 1, µ is called a probability measure, (X, A, µ) is
called a probability space); it S
is called σ-finite if, and only if, there exists a sequence
(Ai )i∈N in A such that X = ∞ i=1 Ai and µ(Ai ) < ∞ for each i ∈ N.
Remark 1.11. In (1.1), we need to extend addition to [0, ∞]. A countable sum of
summands from [0, ∞] is defined to be ∞ if at least one of the summands equals ∞; if
all summands are finite, then the partial sums either converge to some s ∈ R+
0 or they
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 10
If A = P(X), then counting measure is σ-finite if, and only if, X is countable.
1.3 Extension
The examples at the end of the previous section show that there are many measures that
are simple to define and simple to recognize as measures. However, to accomplish our
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 11
original goal of defining a measure on the Borel subsets of Rn that yields the expected
geometric size of intervals, is not so simple. We will be able to do this using the technique
of extension, which is also useful beyond the application we have in mind. The idea is
to first define a suitable [0, ∞]-valued function on some set E that is not a σ-algebra
(e.g. the set of intervals) and then to extend this function to a measure on the σ-algebra
generated by E. We will extend our functions from so-called semirings to σ-algebras.
Useful structures in between are rings and algebras.
(a) E is called a semiring on X if, and only if, it satisfies the following three conditions:
(i) ∅ ∈ E.
(ii) If A, B ∈ E, then A ∩ B ∈ E.
(iii) If A, B ∈ E, then there exist disjoint E1 , . . . , En ∈ E, n ∈ N, such that
n
[
A\B = Ei . (1.2)
i=1
(b) E is called a ring on X if, and only if, it satisfies the following three conditions:
(i) ∅ ∈ E.
(ii) If A, B ∈ E, then A ∪ B ∈ E.
(iii) If A, B ∈ E, then A \ B ∈ E.
where everything is immediate from the respective definitions, except that Def. 1.13(b)
implies Def. 1.13(a)(ii). However, if E is a ring and A, B ∈ E, then
A ∩ B = A \ (A \ B) ∈ E. (1.3)
We also note that E is a ring in the sense of Def. 1.13(b) if, and only if, E is a ring in the
sense of algebra, provided one uses the symmetric difference A ∆ B := (A \ B) ∪ (B \ A)
as addition and ∩ as multiplication (see Appendix B).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 12
Remark and Definition 1.15. Analogous to Prop. 1.5, arbitrary intersections of rings
yield again rings, arbitrary intersections of algebras yield again algebras: Let X T
be a set,
let I be a nonempty index set, and let (Ri )i∈I be a family of rings on X, R := i∈I Ri .
Since ∅ is in each Ri , it is also in R. If A, B ∈ R, then A, B ∈ Ri for each i ∈ I, i.e.
A \ B ∈ Ri and A ∪ B ∈ Ri for each i ∈ I, implying A \ B ∈ R and A ∪ BT∈ A, showing
R to be a ring on X. Now let (Ai )i∈I be a family of algebras on X, A := i∈I Ai . Since
X is in each Ai , it is also in A. Since each Ai is a ring on X, so is A, showing A is an
algebra on X. Thus, if X is a set and E ⊆ P(X), then we can let ρ(E) := ρX (E) denote
the intersection of all rings on X that are supersets of E, and we can let α(E) := αX (E)
denote the intersection of all algebras on X that are supersets of E. We then know ρ(E)
to be a ring and α(E) to be an algebra, namely the smallest ring (resp. algebra) on X
containing E. Thus, we call ρ(E) the ring (and α(E) the algebra) generated by E. We
point out that there exists a (rough) conceptual analogy between the way a σ-algebra
can be built from a semiring and the way a topology can be built from a subbase.
as well as
]a, b] \ ]c, d] = ]a, c] ∪˙ ]d, b] .
|{z} |{z}
= ∅ for c ≤ a = ∅ for b ≤ d
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 13
Moreover,
(A1 × B1 ) \ (A2 × B2 ) = (A1 \ A2 ) × B1 ∪ A1 × (B1 \ B2 )
= (A1 \ A2 ) × B1 ∪˙ (A1 ∩ A2 ) × (B1 \ B2 ) . (1.4)
Using this in (1.4), we can write (A1 × B1 ) \ (A2 × B2 ) as a disjoint union of sets in U,
namely
[M [N
˙ ˙
(A1 × B1 ) \ (A2 × B2 ) = (Ci × B1 ) ∪ ˙ (A1 ∩ A2 ) × Di ,
i=1 i=1
(c) For each n ∈ N, one has ρ(I n ) = IΣn and ρ(IQn ) = IΣ,Q
n
, where I n , IΣn , IQn , and IΣ,Q
n
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 14
Proof. (a) is proved via induction on N . The base case (N = 1) holds, since S is
a semiring. For the induction step, let N ∈ N, A, B1 , . . . , BN , BN +1 ∈ S. By the
induction hypothesis, there exist disjoint C1 , . . . , CM ∈ S such that (1.5) holds true.
Then
N +1 N
! [M
[ [ ˙ [˙ M
A\ Bi = A \ Bi \ BN +1 = Ci \ BN +1 = (Ci \ BN +1 ).
i=1 i=1
i=1 i=1
Since each of the Ci \ BN +1 is a disjoint finite union of sets from S, the induction is
complete.
(b): Since rings are closed under finite unions, R ⊆ ρ(S) is clear. For the remaining
inclusion, it suffices to show that R is a ring. Since ∅ ∈ S S, ∅ ∈ R. LetSM, N ∈ N,
A1 , . . . , AM ∈ S disjoint, B1 , . . . , BN ∈ S disjoint, A := Mi=1 Ai , B :=
N
i=1 Bi . We
have to show A \ B ∈ R and A ∪ B ∈ R. It is
[ [
˙ M ˙ M
A\B = Ai \ Bj .
i=1 j=1
ṠM
Since, by (a), each of the Ai \ j=1 Bj is a disjoint finite union of sets from S, we have
A \ B ∈ R. We also have
[
˙ M [
˙ N
A∩B = Ai ∩ Bj ∈ R,
i=1 j=1
˙ \A) ∪(A∩B)
since each Ai ∩Bj ∈ S by Def. 1.13(a)(ii). Thus, A∪B = (A\B) ∪(B ˙ ∈ R,
completing the proof that R is a ring and the proof of (b).
(c) follows by combining Prop. 1.17(b) with (b).
(a) A map µ : S −→ [0, ∞] is called a content if, and only if, µ satisfies the following
conditions (i) and (ii):
(i) µ(∅) = 0.
(ii) µ is finitely
S additive, i.e., if n ∈ N and A1 , . . . , An ∈ S are disjoint sets such
that ni=1 Ai ∈ S, then
n
! n
[ X
µ Ai = µ(Ai ). (1.7)
i=1 i=1
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 15
The content µ is called finite or bounded if, and only if, µ(A) < ∞ for each A ∈ S;
it is called
S σ-finite if, and only if, there exists a sequence (Ai )i∈N in S such that
X= ∞ i=1 i and µ(Ai ) < ∞ for each i ∈ N.
A
(b) A content µ : S −→ [0, ∞] is called a premeasure if, and only if, it is countably
additive (also called σ-additive),
S i.e., if (Ai )i∈N is a sequence in S consisting of
disjoint sets such that ∞ A
i=1 i ∈ S, then
∞
! ∞
[ X
µ Ai = µ(Ai ).
i=1 i=1
(b) Among the most important premeasures is so-called Lebesgue premeasure λn , which,
for n ∈ N, is defined on the semiring I n on Rn from Ex. 1.8(c) by letting
(Q
n n
n n + n j=1 (bj − aj ) for I =]a, b], a, b ∈ R , a < b,
λ : I −→ R0 , λ (I) :=
0 for I = ∅.
We will study it in detail in Sec. 1.3.3 below, where we will show in Th. 1.31 that
it is, indeed, a premeasure.
(c) The following example shows that not every content is a premeasure: Consider
S := {]a, b] : a, b ∈ R, 0 ≤ a ≤ b ≤ 1}. Then S is clearly a semiring on ]0, 1] (cf.
Ex. 1.16(b)). Define
0 for I = ∅,
µ : S −→ [0, ∞], µ(I) := b − a for I =]a, b], 0 < a < b ≤ 1,
∞ for I =]0, b], 0 < b ≤ 1.
and
n
X n
X
µ ]0, s0 ] + µ ]si−1 , si ] = ∞ + (si − si−1 )
∀ ∀ i=1 i=1
n∈N 0<s0 <s1 <···<sn ≤1
= ∞ = µ ]0, sn ] .
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 16
1
However, µ is not a premeasure, since, letting, for each n ∈ N, In :=] n+1 , n1 ], one
Ṡ∞
has ]0, 1] = n=1 In , but
X∞ X∞
1 1 1
µ(In ) = − = lim 1 − = 1 6= µ ]0, 1] = ∞.
n=1 n=1
n n+1 n→∞ n+1
Before compiling important properties of contents and premeasures in Th. 1.22 and Th.
1.24 below, we show in the following Prop. 1.21 that each content on a semiring extends
in a unique way to a content on the generated ring.
Proposition 1.21. Let S be a semiring on the set X and let µ : S −→ [0, ∞] be a
content. Furthermore, let R := ρ(S). Then there is a unique extension ν : R −→ [0, ∞]
of µ such that ν is a content. Moreover ν is a premeasure if, and only if, µ is a
premeasure.
proving that ν is well-defined. Then ν↾S = µ is also clear. To prove that ν is a content,
we still need to check that it is finitely additive. To this end, let A, B ∈ R be disjoint.
S
Then there are disjoint A1 , . . . , AN , B1 , . . . , BM ∈ S; M, N ∈ N, such that A = Ni=1 Ai
SM
and B = i=1 Bi . In consequence,
N
X M
X
ν(A ∪ B) = µ(Ai ) + µ(Bi ) = ν(A) + ν(B),
i=1 i=1
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 17
as needed.
Since ν↾S = µ, if ν is a premeasure, so is µ. For the converse, assume µSto be a premeasure
and let (Ai )i∈N be a sequence of disjoint sets in R such that A := ∞ i=1 Ai ∈ R. Then
SN
there are disjoint sets B1 , . . . , BN ∈ S, N ∈ N, with A = i=1 Bi , and, for each i ∈ N,
S i
there are disjoint sets Ci1 , . . . , CiNi ∈ S, Ni ∈ N, with Ai = Nj=1 Cij . Then
[
˙ ∞ ˙ ∞ [
[ ˙ Nj
∀ Bi = Bi ∩ A = (Bi ∩ Aj ) = (Bi ∩ Cjk ).
i∈{1,...,N } j=1 j=1 k=1 | {z }
∈S
Thus, as µ is a premeasure,
Nj
∞ X ∞
X X
∀ µ(Bi ) = µ(Bi ∩ Cjk ) = ν(Bi ∩ Aj ),
i∈{1,...,N }
j=1 k=1 j=1
implying
N
X ∞
N X
X ∞
X ∞
X
ν(A) = µ(Bi ) = ν(Bi ∩ Aj ) = ν(A ∩ Aj ) = ν(Aj ),
i=1 i=1 j=1 j=1 j=1
proving ν to be a premeasure.
During the construction of a content µ, it can be helpful to first define and study it on
a semiring. However, Prop. 1.21 tells us that we may assume µ to be defined on a ring
when considering a content’s general properties, as in the following Th. 1.22 (one still
has to use some care, though, cf. Ex. 1.25(c) below).
Theorem 1.22. Let R be a ring on the set X and let µ : R −→ [0, ∞] be a content.
Then the following rules hold, where we assume A, B ∈ R as well as Ai ∈ R for each
i ∈ N.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 18
˙
Proof. (a): If B ⊆ A, then A = B ∪(A \ B), implying
If µ(A∪B) < ∞, then we add µ(A) = µ(A∪B)−µ(B \A) to the first equation to obtain
(c). If µ(A ∪ B) = µ(A) = ∞, then (c) also holds. In the remaining case, µ(A ∪ B) = ∞
and µ(A) < ∞, i.e. µ(B \ A) = ∞, implying µ(B) = ∞, i.e. (c) holds once again.
Pn
(d) holds, since, by finite additivity and monotonicity, i=1 µ(Ai ) ≤ µ(A) holds for
each n ∈ N.
S Ṡn Sn−1
(e): Since ni=1 Ai = i=1 (Ai \ j=1 Aj ), we have
n
! n n−1
! n
[ X [ X
µ Ai = µ Ai \ Aj ≤ µ(Ai ).
i=1 i=1 j=1 i=1
Ṡn Sn−1
(f) follows with the same kind of trick as in (e): Since A = i=1 (A ∩ (Ai \ j=1 Aj )),
we have !!
X∞ n−1
[ ∞
X
µ(A) = µ A ∩ Ai \ Aj ≤ µ(Ai ),
i=1 j=1 i=1
Theorem 1.24. Let R be a ring on the set X, let µ : R −→ [0, ∞] be a content, and
consider the following statements:
(i) µ is a premeasure.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 19
(ii) µ is continuous from below, i.e. if A ∈ R and (An )n∈N is a sequence in R, then
An ↑ A ⇒ lim µ(An ) = µ(A). (1.9)
n→∞
(iii) µ is continuous from above, i.e. if A ∈ R and (An )n∈N is a sequence in R, then
µ(A1 ) < ∞ ∧ An ↓ A ⇒ lim µ(An ) = µ(A). (1.10)
n→∞
Proof. (i) ⇒ (ii): Let A ∈ R and let (An )n∈N be a sequence in R such that An ↑ A.
Ṡ∞
Then A = A1 ∪˙ i=2 (Ai \ Ai−1 ), implying
∞ n
!
X X
µ(A) = µ(A1 ) + µ(Ai \ Ai−1 ) = lim µ(A1 ) + µ(Ai \ Ai−1 ) = lim µ(An ),
n→∞ n→∞
i=2 i=2
as desired.
⇒ A1 \ An ↑ A1 \ A.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 20
proving (iii).
(iii) ⇒ (iv) is immediate by setting A := ∅.
(iv) ⇒ (iii): Let A ∈ R and let (An )n∈N be a sequence in R such that An ↓ A, µ(A1 ) <
∞. For each n ∈ N, set Bn := An \ A ∈ R. Then Bn ↓ ∅ and µ(B1 ) ≤ µ(A1 ) < ∞, i.e.
we can apply (iv) to obtain
lim µ(An ) − µ(A) = lim µ(Bn ) = 0,
n→∞ n→∞
proving (iii).
Finally, assume µ to be finite and (iv). Let A ∈ R and let (An )n∈N be a sequence in R
such that An ↑ A. Then A \ An ↓ ∅ and, since µ(A \ A1 ) < ∞, (iv) yields
lim µ(A) − µ(An ) = lim µ(A \ An ) = 0,
n→∞ n→∞
The following Ex. 1.25(a)–(c) provides some counterexamples related to Th. 1.24 above.
Example 1.25. (a) The following example shows that one cannot omit the condition
µ(A1 ) < ∞ in (iii) and (iv) of Th. 1.24: Consider the counting measure of Ex.
1.12(b) with X := N. If, for each n ∈ N, An := {k ∈ N : k ≥ n}. Then An ↓ ∅, but
µ(An ) = ∞ for each n ∈ N.
(b) The following example shows statements (iii) and (iv) of Th. 1.24 are, in general,
not equivalent to statements (i) and (ii): Let X be a countable infinite set (e.g.,
X = N). Then A := {A ⊆ X : A finite or Ac finite} is clearly an algebra on X and
(
0 for A finite,
µ : A −→ [0, ∞], µ(A) :=
∞ for A infinite,
clearly defines a content on A. Moreover, µ satisfies (iii) and (iv) of Th. 1.24, since,
if A ∈ A and (An )n∈N is a sequence in A with An ↓ A, then µ(A1 ) < ∞ implies that
all AnP and A are finite with µ(An ) = µ(A) = 0. However, µ is not a premeasure,
since x∈X µ({x}) = 0 6= ∞ = µ(X) (where we used that X is the countable
disjoint union of the {x}).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 21
(c) The following example shows Th. 1.24 does, in general, not hold if the ring R is
replaced with a semiring S: Let X := Q and
S := I 1 |Q := {]a, b] ∩ Q : a, b ∈ R, a ≤ b}.
defines a finite content, satisfying Th. 1.24(ii) on S (and, thus, also Th. 1.24(iii),(iv)
on S, since the corresponding proofs of Th. 1.24 remain valid), but (the extension of)
µ does not satisfy Th. 1.24(ii) on R := ρ(S) (in particular, µ is not a premeasure):
Since
n
X n
X
∀ ∀ µ ]si−1 , si ] ∩ Q = (si − si−1 ) = sn − s0 = µ ]s0 , sn ] ∩ Q ,
n∈N s0 <s1 <···<sn
i=1 i=1
showing Th. 1.24(ii) holds on S. Next, we show that (the extension of) µ does
not satisfy Th. 1.24(ii) on R := ρ(S): Let A :=]0, 1] ∩ Q and let q1 , q2 , . . . be
an enumeration of the (countable) set A, 0 < ǫ < 1. We recursively construct a
)n∈N , An =]an , bn ] ∩ Q, of disjoint sets in S, and a sequence (Bn )n∈N
sequence (An S
in R, Bn := ni=1 Ai , by letting a1 := max{0, q1 − ǫ 2−1 }, b1 := q1 (note µ(B1 ) =
µ(A1 ) ≤ ǫ 2−1 ) and, for n > 1, let Nn := min{k ∈ N : qk ∈ / Bn−1 }, bn := qNn ,
an := max {t ∈ Bn−1 : t < bn } ∪ {bn − ǫ 2 } , noting µ(An ) ≤ ǫ 2−n and
−n
n
X ∞
X
µ(Bn ) = µ(Ai ) < ǫ 2−i = ǫ < 1. (1.12)
i=1 i=1
Ṡn−1
We
Pn−1 point out that N n above is well-defined, since, if B n−1 = i=1 Ai = A, then
i=1 µ(Ai ) < ǫ < 1 = µ(A) would mean that µ were not a content. Moreover,
since Bn ↑ A (note that every qk must be in precisely one of the Ai ), (1.12) shows
that µ does not satisfy Th. 1.24(ii) on R.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 22
At this point, we are in a position to perform the first step of the process described
at the beginning of Sec. 1.3: We can define a premeasure on a semiring that, when
extended to all Borel sets in Rn , will satisfy the conditions of Def. 1.1 (cf. Ex. 1.47).
Definition 1.26. Let n ∈ N and recall the definition of the semiring I n on Rn from
Ex. 1.8(c). The function
(Q
n n
j=1 (bj − aj ) for I =]a, b], a, b ∈ R , a < b,
λn : I n −→ R+ 0 , λ n
(I) := (1.13)
0 for I = ∅,
The main objective of this section is to show that λn is, indeed, a premeasure. It turns
out not to be as easy as one might have thought. We will first prove it to be a content in
Prop. 1.28 and then to be a premeasure in Th. 1.31. We introduce the following notions
that will be useful to prove Prop. 1.28:
(b) Given a half-open interval I =]a, b] ∈ I n , a, b ∈ Rn with a < b, we call the set
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 23
We say that the hyperplanes Hk1 ,α1 , . . . , HkM ,αM cut the interval I into the intervals
I1 , . . . , IN if, and only if, the following two conditions hold:
+ −
∀ ∀ Ii ⊆ H kj ,αj ∨ Ii ⊆ H kj ,αj , (1.15a)
i∈{1,...,N } j∈{1,...,M }
( )
n o [N
(kj , αj ) : j ∈ {1, . . . , M } = (k, α) ∈ C(Ij ) : ak < α < bk (1.15b)
j=1
(condition (1.15a) says that each small interval Ij has to lie on precisely one side of
each of the cutting halfspaces; condition (1.15b) says that all halfspaces intersect
the interior of I and that the cutting halfspaces extend all the faces of the Ij that
intersect the interior of I to the boundary of I).
Proof. We conduct the proof in several steps. First, we consider an interval I =]a, b],
a, b ∈ Rn , a < b, that is cut by a hyperplane Hk,α , k ∈ {1, . . . , n} and α ∈]ak , bk [, into
intervals I1 , I2 ∈ I n in the sense of Def. 1.27. Then, clearly I1 ∪˙ I2 , where I1 =]a1 , b1 ],
I2 =]a2 , b2 ], with
Moreover,
n k−1
! n
!
Y Y Y
λn (I) = (bj − aj ) = (bj − aj ) (α − ak ) (bj − aj )
j=1 j=1 j=k+1
k−1
! n
!
Y Y
+ (bj − aj ) (bk − α) (bj − aj ) = λn (I1 ) + λn (I2 ).
j=1 j=k+1
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 24
we have
N
X N X
X
n n
λ (I) = λ (Ii ) = λn (Ik ),
i=1 i=1 k∈Ji
completing the induction. Finally, given I =]a, b] ∈ I n (a, b ∈ Rn with a < b), let
ṠK
K ∈ N, and let I1 , . . . , IK ∈ I n be disjoint intervals such that I = j=1 Ij . Set
( K
)
[
J := (k, α) ∈ C(Ij ) : ak < α < bk ,
j=1
∀ ∃ Ai ⊆ Ij
i∈{1,...,N } j∈{1,...,K}
and, thus,
[˙ K j
Ai : i ∈ {1, . . . , N } = Aι : ι ∈ {1, . . . , Nj } .
j=1
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 25
Proposition 1.30. Let (X, T ) be a topological space, let S be a semiring on X, and let
µ : S −→ R+
0 be an inner regular content. Then the following assertions hold true:
Proof. (a): According to Prop. 1.18(b), R consists of disjoint finite unions of sets from
S. Thus, if µ is finite Son S, then µ is also finite on R. Now let A1 , . . . , An ∈ S be
disjoint, n ∈ N, A := ni=1 Ai . Given ǫ ∈ R+ , for each i ∈ {1, . . . , n}, S
let Ki ∈ S be
such that K i is compact, K i ⊆ Ai , and µ(Ai ) ≤ µ(Ki ) + n . Let K := ni=1 Ki . Then
ǫ
S
K ∈ R, K = ni=1 K i is compact, K ⊆ A, and
n
ǫ
n
X X
µ(A) = µ(Ai ) ≤ µ(Ki ) + = µ(K) + ǫ,
i=1 i=1
n
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 26
T
then H n = ni=1 H T∞ n ∈ N and, by the finite intersection property of the
Ti∞6= ∅ for each
compact set H 1 , n=1 An ⊇ n=1 H n 6= ∅ as desired. It remains to show (1.17). We
accomplish this via proving, for each n ∈ N,
µ(Hn ) ≥ µ(An ) − ǫ(1 − 2−n ) ≥ ǫ − ǫ + ǫ 2−n = ǫ 2−n > 0 (1.18)
by induction on n. For n = 1, H1 = K1 and (1.18) holds, since µ(A1 ) ≤ µ(K1 ) + 2−1 ǫ
by the choice of K1 . For the induction step, fix n ∈ N and estimate
Th. 1.22(c)
µ(Hn+1 ) = µ(Hn ∩ Kn+1 ) = µ(Kn+1 ) + µ(Hn ) − µ(Hn ∪ Kn+1 )
ind.hyp.
≥ µ(Kn+1 ) + µ(An ) − ǫ(1 − 2−n ) − µ(Hn ∪ Kn+1 ). (1.19)
Due to the choice of Kn+1 , we have µ(Kn+1 ) ≥ µ(An+1 )−2−(n+1) ǫ as well as Kn+1 ∪Hn ⊆
An+1 ∪ An = An and, thus, µ(Kn+1 ∪ Hn ) ≤ µ(An ), which we use in (1.19) to obtain
µ(Hn+1 ) ≥ µ(An+1 ) − 2−(n+1) ǫ − ǫ(1 − 2−n ) = µ(An+1 ) − ǫ(1 − 2−(n+1) ),
completing the induction as well as the proof of the proposition.
Theorem 1.31. For each n ∈ N, Lebesgue premeasure λn as defined in Def. 1.26
constitutes a premeasure on the semiring I n .
Proof. We know λn to be a content from Prop. 1.28. Thus, the assertion will follow from
Prop. 1.30, if we can show λn to be inner regular with respect to the norm topology
on Rn . To this end, let a, b, c ∈ Rn , a < c < b, A =]a, b], K =]c, b]. Then K ⊆ A is
compact and, letting Mn := max{bj − aj : j ∈ {1, . . . , n}},
n
X
n n
λ (A) ≤ λ (K) + Mnn−1 (cj − aj ). (1.20)
j=1
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 27
as needed to conclude the induction. Now fix n ∈ N, ǫ > 0, and let A and M := Mn
b +a
be as above. For each j ∈ {1, . . . , n}, let cj := min{ j 2 j , aj + nMǫn−1 }, K :=]c, b] (as
before). Then
n
X
(1.20) ǫ
λn (A) ≤ λn (K) + M n−1 (cj − aj ) ≤ λn (K) + M n−1 n n−1
= λn (K) + ǫ,
j=1
nM
(i) η(∅) = 0.
(ii) Monotonicity:
∀ η(A) ≤ η(B). (1.21)
A⊆B⊆X
(iii) η is countably subadditive (also called σ-subadditive), i.e., if (Ai )i∈N is a sequence
in P(X), then !
∞
[ X∞
η Ai ≤ η(Ai ). (1.22)
i=1 i=1
Remark 1.33. Since we know measures to be monotone and σ-subadditive, every mea-
sure on P(X) is an outer measure. However, the converse is not true in general: For
example, (
0 if A = ∅,
η : P(X) −→ [0, ∞], η(A) :=
1 if A 6= ∅,
defines an outer measure that is not a measure if X has more than one element.
Definition 1.34. Let X be a set and let η : P(X) −→ [0, ∞] be an outer measure on
X. We call A ⊆ X η-measurable if, and only if,
Lemma 1.35. Let X be a set and let η : P(X) −→ [0, ∞] be an outer measure on X,
A ⊆ X.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 28
Proof. (a) If η(A) = 0, then the monotonicity of η implies, for each Q ⊆ X, η(Q∩A) = 0
and, thus η(Q ∩ A) + η(Q ∩ Ac ) = η(Q ∩ Ac ) ≤ η(Q), showing A to be η-measurable.
Analogously, we see A to be η-measurable for η(Ac ) = 0.
(b) is clear, since (1.23) always holds for η(Q) = ∞.
(c) is also clear since ≤ always holds in (1.23) by the subadditivity of η.
Proposition 1.36. Let X be a set and let η : P(X) −→ [0, ∞] be an outer measure on
X. Define
Aη := {A ⊆ X : A is η-measurable}. (1.25)
Then Aη is a σ-algebra and η↾Aη is a measure.
proving A ∪ B ∈ Aη . Now
establishing Aη to be an algebra.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 29
S∞
Now let (Ai )i∈N be a sequence of disjoint sets in Aη and let A := i=1 Ai . We claim
that A ∈ Aη and
X∞
η(A) = η(Ai ). (1.27)
i=1
via induction
S on n: For n = 1, there is nothing to prove. For n >
Sn−1 1, we apply (1.24)
Sn−1
with Q ∩ ni=1 Ai instead of Q and i=1 Ai instead of A (we know i=1 Ai ∈ Aη , as Aη
is an algebra) to obtain
n
! n−1
! n
[ [ ind.hyp.
X
η Q∩ Ai = η Q ∩ Ai + η(Q ∩ An ) = η(Q ∩ Ai )
i=1 i=1 i=1
S
as needed. Now we prove A ∈ Aη and (1.27): For each n ∈ N, since ni=1 Ai ∈ Aη , we
can apply (1.28) to estimate
n
! n
!c ! n
(1.23) [ [ (1.28),Def. 1.32(ii) X
η(Q) ≥ η Q ∩ Ai + η Q ∩ Ai ≥ η(Q ∩ Ai ) + η(Q ∩ Ac )
i=1 i=1 i=1
implying
∞
X Def. 1.32(iii) Def. 1.32(iii)
η(Q) ≥ η(Q∩Ai )+η(Q∩Ac ) ≥ η(Q∩A)+η(Q∩Ac ) ≥ η(Q). (1.29)
i=1
Thus, all terms in (1.29) must be equal, proving A ∈ Aη and, for Q := A, (1.27).
Due to (1.27), η↾Aη is a measure, provided Aη is a σ-algebra. So, finally,
S let (Ai )i∈N be
a sequence in Aη (now the Ai do not have to be disjoint) and let A := ∞ i=1 Ai . Then
i−1
!
[˙ ∞ [
A= Ai \ Ak ∈ Aη
i=1
k=1
| {z }
∈Aη
The following Prop. 1.37 yields a useful construction method for outer measures:
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 30
defines an outer measure on X. This outer measure is called the outer measure induced
by µ.
Since (1.31) tivially holds if, for at least one i ∈ N, η(Ai ) = ∞, we may assume
η(Ai ) < ∞ for each i ∈ N. Let ǫ ∈ R+ . Then, for each i ∈ N, there exists a sequence
(Bik )k∈N in S such that
∞
[ ∞
X
Ai ⊆ Bik ∧ µ(Bik ) < η(Ai ) + ǫ · 2−i .
k=1 k=1
S
Since (Bik )(i,k)∈N2 is a countable family (i.e. a sequence) in S with A ⊆ (i,k)∈N2 Bik ,
∞ X
X ∞ ∞
X ∞
X
η(A) ≤ µ(Bik ) ≤ η(Ai ) + ǫ · 2−i = ǫ + η(Ai )
i=1 k=1 i=1 i=1
(b) If µ is a premeasure, then η↾S = µ (thus, in this case, η↾Aη is a measure that extends
the premeasure µ to a σ-algebra containing σ(S)).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 31
Proof. (a): According to Prop. 1.37, η is an outer measure. It remains to show that
each A ∈ S is η-measurable, i.e. we have to verify (1.23), given some A ∈ S. It will be
convenient to extend the content µ to a content on R := ρ(S) via Prop. 1.21 (we still
call the extension µ). Then (1.30) and Prop. 1.18(b) imply
(∞ ∞
)
X [
∀ η(Q) = inf µ(Ai ) : (Ai )i∈N sequence in R, Q ⊆ Ai . (1.33)
Q∈P(X)
i=1 i=1
Now taking the infimum in (1.34) over all admissible sequences (Ai )i∈N in S according
to (1.30), we obtain η(Q) ≥ η(Q ∩ A) + η(Q ∩ Ac ), proving S ⊆ Aη as claimed.
(b): The inequality η ↾S ≤ µ is immediate from (1.30). It remains to show µ ≤ η ↾S .
We know from Prop. 1.21 that, if µ is a premeasure on S, then the S extended µ is a
∞
premeasure on R. Thus, if (Ai )i∈N
P∞is a sequence in R such that Q ⊆ i=1 Ai , Q ∈ S,
then, by Th. 1.22(f), µ(Q) ≤ i=1 µ(Ai ). Taking the infimum over all admissible
sequences according to (1.33) yields µ(Q) ≤ η(Q) and µ ≤ η↾S as needed.
(c): If µ is not a premeasure on S, S∞then there exists AP∈∞S and a sequence (Ai )i∈N of
disjoint sets in S suchPthat A = i=1 Ai and µ(A) 6= i=1 µ(Ai ). Due to Th. 1.22(d)
(applied to µ on R), ∞ i=1 µ(Ai ) ≤ µ(A). Thus,
∞
X
η(A) ≤ µ(Ai ) < µ(A),
i=1
The following simple Ex. 1.39 shows that, in general, one can not expect the extension
of a premeasure µ on a semiring S to a measure on σ(S) to be unique:
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 32
Example 1.39. Let X be a nonempty set and R := {∅}. Then R is a semiring (and
even a ring) on X, and µ : R −→ [0, ∞], µ(∅) := 0, defines a premeasure. Then
A := σ(R) = {∅, X} and
(
0 for A = ∅,
∀ µα : A −→ [0, ∞], µα (A) :=
α∈[0,∞] α for A = X,
The problem in Ex. 1.39, as it turns out, is that the premeasure µ on the semiring S is
not σ-finite. In preparation for the proof of the uniqueness Th. 1.45, we develop some
additional technical tools that are also of measure-theoretic use beyond the application
in the present section.
∀ A ∩ B ∈ E.
A,B∈E
(b) E is called a monotone class on X if, and only if, it satisfies the following two
conditions:
(c) E is called a Dynkin system on X if, and only if, it satisfies the following three
conditions:
(i) X ∈ E.
(ii) If A ∈ E, then Ac ∈ E.
S
(iii) If (En )n∈N is a sequence of disjoint sets in E, then n∈N En ∈ E.
Remark and Definition 1.41. Arbitrary intersections of monotone classes yield again
monotone classes, arbitrary intersections of Dynkin systems yield again Dynkin systems:
Let X be a set, let I T
be a nonempty index set, and let (Mi )i∈I be a family of monotone
classes on X, M := i∈I Mi . If A ⊆ X and (En )n∈N is a sequence in M with En ↑ A
(resp. En ↓ A), then, for each i ∈ I, (En )n∈N is a sequence in Mi , i.e. A ∈ Mi for each
i ∈ I, proving M to be a monotone class on X. Now let (Di )i∈I be a family of Dynkin
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 33
T
systems on X, D := i∈I Di . Since X is in each Di , it is also in D. If A ∈ D, then
A ∈ Di for each i ∈ I, i.e. Ac ∈ Di for each i ∈ I, implying Ac ∈ D. If (En )n∈N is
S sets in D then, for each i ∈ I, (En )n∈N is a sequence of disjoint
a sequence of disjoint
sets in Di . Thus, n∈N En ∈ D, showing D is a Dynkin system on X. Thus, if X
is a set and E ⊆ P(X), then we can let µ(E) := µX (E) denote the intersection of all
monotone classes on X that are supersets of E, and we can let δ(E) := δX (E) denote the
intersection of all Dynkin systems on X that are supersets of E. We then know µ(E) to
be a monotone class and δ(E) to be a Dynkin system, namely the smallest monotone
class (resp. Dynkin system) on X containing E. Thus, we call µ(E) the monotone class
(and δ(E) the Dynkin system) generated by E.
Proposition 1.42. Let X be a set, E ⊆ P(X). Then E is a Dynkin system on X if,
and only if, the following conditions (i) – (iii) hold:
(i) X ∈ E.
Proof. Assume (i) – (iii) hold. We have to show E is closed under complements and
under countable unions of disjoint sets. If B ∈ E, then, choosing A := X in (ii) shows
B c = X\B ∈ E. If (Ei )i∈N is a sequence of disjoint sets in E, then E1 ∪E2 = (E1c \E2 )c ∈ E
c c
Snthen E2 ⊆ E1 and E1 \ E2 ∈ E by (ii)). Now
(if E1 , E2 are disjoint, S∞an induction over
n ∈ N shows An := i=1 Ei ∈ E for each n ∈ N. Since An ↑ A := i=1 Ei , (iii) implies
A ∈ E, showing E to be a Dynkin system.
Conversely, assume E to be a Dynkin system. We have to show that (ii) and (iii) hold.
If A, B ∈ E with B ⊆ A, then Ac , B, ∅ are disjoint sets in E, yielding
C := Ac ∪ B ∪ ∅ ∪ · · · ∈ E, A \ B = C c ∈ E,
showing E satisfies (ii). Now let A ⊆ X and let (Ei )i∈N be a sequence of sets in E.
Ṡ∞ S
If Ei ↑ A, then A = E1 ∪˙ i=2 (Ei \ Ei−1 ) ∈ E. If Ei ↓ A, then Ac = ∞ c
i=1 Ei , i.e.
Eic ↑ Ac ∈ E. Thus, A ∈ E as well and E is a monotone class, which completes the
proof.
Proposition 1.43. Let X be a set and let E ⊆ P(X) be a Dynkin system on X. Then
the following statements are equivalent:
(i) E is ∩-stable.
(ii) E is a σ-algebra.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 34
Proof. As every σ-algebra is ∩-stable, it only remains to show (i) implies (ii). If A, B ∈ E
and E is ∩-stable, then A ∪ B = (A ∩ B c ) ∪(B˙ ˙
∩ Ac ) ∪(A ∩SB) ∈ E. Let (Ei )i∈N be a
n
sequence of sets in
S∞E. An induction over n ∈ N shows A n := i=1 Ei ∈ E for each n ∈ N.
Since An ↑ A := i=1 Ei and the Dynkin system E is a monotone class, A ∈ E, showing
E to be a σ-algebra.
Proposition 1.44. Let X be a set and let E ⊆ P(X) be ∩-stable. Then δ(E) = σ(E).
Proof. Since every σ-algebra is a Dynkin system, we have D := δ(E) ⊆ σ(E). So it only
remains to show that D is a σ-algebra. Moreover, according to Prop. 1.43, it suffices to
show D is ∩-stable. We will achieve this by showing
While this strategy might seem strange at first, in various variants, it turns out to be
quite powerful in measure theory, and we will use it again in the proof of Th. 1.45
below. We note that, for each D ∈ D, D(D) is a Dynkin system: X ∈ D(D), since
X ∩ D = D ∈ D. If A ∈ D(D), then Ac ∈ D(D), as Ac ∩ D = D \ (A ∩ S D) ∈ D by Prop.
1.42(ii). If (Ai )i∈N is a sequence of disjoint sets in D(D), then A := ∞ i=1 Ai ∈ D(D),
since [ ˙ ∞
A∩D = (Ai ∩ D) ∈ D.
i=1
Since E is ∩-stable, we have E ⊆ D(E) for each E ∈ E. Thus, D = δ(E) ⊆ D(E) for each
E ∈ E, since D(E) is a Dynkin system. Thus, if E ∈ E and D ∈ D, then D ∈ D(E), i.e.
D ∩ E ∈ D. But this means also E ∈ D(D), showing E ⊆ D(D), which proves (1.35)
and the proposition.
Theorem 1.45. Let (X, A) be a measurable space and let µ, ν : A −→ [0, ∞] be mea-
sures. Moreover, let E ⊆ P(X) have the following properties:
(i) σ(E) = A.
(iii) E is ∩-stable.
Then µ = ν.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 35
First, we note that, for each E ∈ E with µ(E) = ν(E) < ∞, D(E) is a Dynkin system:
X ∈ D(E), since X ∩ E = E ∈ E. If A ∈ D(E), then Ac ∈ D(E), since
(iii)
µ(Ac ∩ E) = µ E \ (A ∩ E) = µ(E) − µ(A ∩ E) = ν(E) − ν(A ∩ E)
= ν E \ (A ∩ E) = ν(Ac ∩ E).
S
If (Ai )i∈N is a sequence of disjoint sets in D(E), then A := ∞ i=1 Ai ∈ D(E), since
[∞ ∞
X ∞
X
˙
µ(A ∩ E) = µ (Ai ∩ E) = µ(Ai ∩ E) = ν(Ai ∩ E)
i=1
i=1 i=1
[∞
˙
=ν (Ai ∩ E) = ν(A ∩ E).
i=1
proving (1.36). Let (Ei )i∈N be as in (iv) and define, F0 := ∅ as well as, for each n ∈ N,
S Ṡn
Fn := ni=1 Ei . Then Fn ↑ X. For each n ∈ N, we have Fn = i=1 (Ei ∩ Fi−1 c
), implying
n
X n
X
c (1.36) c
∀ µ(A ∩ Fn ) = µ(A ∩ Fi−1 ∩Ei ) = ν(A ∩ Fi−1 ∩Ei ) = ν(A ∩ Fn )
n∈N,
i=1
| {z } i=1
| {z }
A∈A ∈A ∈A
and
∀ µ(A) = lim µ(A ∩ Fn ) = lim ν(A ∩ Fn ) = ν(A),
A∈A n→∞ n→∞
(b) One can strengthen (a) to the following comparison result: If α, β : σ(S) −→ [0, ∞]
are measures that satisfy α↾S ≤ β↾S , where β↾S is σ-finite, then α ≤ β holds on all
of σ(S).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 36
Proof. (a): Existence was shown in Th. 1.38(b), whereas uniqueness is due Th. 1.45,
which applies with E := S, since the semiring S is ∩-stable.
(b): Since β ↾S is σ-finite and α ↾S ≤ β ↾S , α ↾S must be σ-finite as well. Let α∗ , β ∗ :
P(X) −→ [0, ∞] be the outer measures induced by α↾S and β↾S , respectively (as defined
in Prop. 1.37). Then the Carathéodory Extension Th. 1.38 combined with (a) yields
α = α∗ ↾σ(S) and β = β ∗ ↾σ(S) . In consequence, α↾S ≤ β ↾S combined with (1.30) proves
α ≤ β.
Example 1.47. Let n ∈ N. From Th. 1.31, we know the Lebesgue premeasure λn to
be a premeasure on the semiring I n on Rn . Clearly, λn is also σ-finite. Thus, according
to Cor. 1.46(a), λn can be uniquely extended to a measure β n on B n = σ(I n ), called
Lebesgue-Borel measure, where B n denotes the σ-algebra of Borel sets in Rn (see Ex.
1.8(c)). According to the Carathéodory Extension Th. 1.38, one obtains β n by first
extending λn to the induced outer measure λn : P(Rn ) −→ [0, ∞] (also called the
Lebesgue outer measure) and restricting it, first to Ln := Aλn and then to B n . The
measure λn : Ln −→ [0, ∞] is called Lebesgue measure. The notation β n will be used
when it is to be emphasized that λn is considered only on B n . The sets in B n are called
β n -measurable or Borel-measurable, the sets in Ln are called λn -measurable or Lebesgue-
measurable. According to Cor. 1.52 below, Lebesgue measure is the so-called completion
of Lebesgue-Borel measure (see Def. 1.48 and Def. 1.50) and one can show that B n ( Ln ,
where B n has the same cardinality as R and Ln has the same cardinality as P(R) (see
[Els07, Sec. 2.8.3] and the last paragraph of Sec. 1.5.3 below). We will reencounter the
important measures β n and λn throughout this class. In particular, they will be seen to
satisfy Def. 1.1, except that they are not defined on all of P(Rn ).
1.3.6 Completion
Sets of measure zero play a special role in measure theory, giving rise to the following
definitions.
Definition 1.48. Let (X, A, µ) be a measure space.
(a) A set N ∈ A is called a µ-null set (or just a null set if µ is understood) if, and only
if, µ(N ) = 0.
(b) (X, A, µ) and µ are called complete if, and only if, each subset of a µ-null set is
µ-measurable, i.e. if, and only if,
∀ ∀ µ(A) = 0 ⇒ B ∈ A .
A∈A B⊆A
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 37
If (X, A, µ) is a measure space, then µ can be used as the content of the Carathéodory
Extension Th. 1.38, which yields the extension η↾Aη . According to Lem. 1.35(a), η↾Aη is
complete. Thus, every measure can be extended to a complete measure. This can also
be seen, using a simpler and more natural construction that does not make use of outer
measures, according to the following Th. 1.49. If µ is σ-finite, then both constructions
yield the same complete extension (the so-called completion of µ), see Th. 1.51 below.
Theorem 1.49. Let (X, A, µ) be a measure space. Define N to be the set consisting of
all subsets of µ-null sets, i.e.
N := N ∈ P(X) : ∃ N ⊆ A ∧ µ(A) = 0 . (1.37a)
A∈A
Also define
à := {A ∪ N : A ∈ A, N ∈ N }, (1.37b)
µ̃ : Ã −→ [0, ∞], µ̃(A ∪ N ) := µ(A) for A ∈ A, N ∈ N . (1.37c)
(a) Ã is a σ-algebra, µ̃ is well-defined by (1.37c), and (X, Ã, µ̃) is complete. Moreover,
µ̃ is the only content on à that is an extension of µ.
(b) Each complete measure ν that extends µ must also extend µ̃ (i.e. µ̃ is the “smallest”
complete extension of µ).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 38
extension of µ. To verify σ-additivity of µ̃, let (Ak )k∈N be a sequence of disjoint sets in
A and (Nk )k∈N a sequence of disjoint sets in N . Then
∞
! ∞
! ∞
!!
[ [ [
µ̃ (Ak ∪ Nk ) = µ̃ Ak ∪ Nk
k=1 k=1 k=1
∞
! ∞ ∞
[ X X
=µ Ak = µ(Ak ) = µ(Ak ∪ Nk ),
k=1 k=1 k=1
] To this end,
Proof. As we know η ↾Aη to be complete, it remains to prove Aη ⊆ σ(S).
let B ∈ Aη . First, we assume η(B) <S∞. Then, forPeach n ∈ N, there exists a sequence
(Ank )k∈N of sets in S such that B ⊆ k=1 Ank and ∞
∞ 1
k=1 µ(Ank ) ≤ η(B) + n . Defining
∞ [
\ ∞
A := Ank ,
n=1 k=1
we obtain
1
A ∈ σ(S), B ⊆ A, . ∀ η(A) ≤ η(B) +
n n∈N
Thus, η(A) = η(B). We can now apply what we have just shown to A \ B instead of
B, proving the existence of C ∈ σ(S) such that A \ B ⊆ C and η(C) = η(A \ B) =
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 39
Definition and Remark 1.54. Let X be a set, and let E ⊆ P(X) be a collection of
subsets. If B ⊆ X, then let E|B := {A ∩ B : A ∈ E} denote the trace of E on B, also
known as the restriction of E to B. Consider the canonical inclusion map f : B −→ X,
f (a) = a. Then,
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 40
∞ ∞
!
[ [
f −1 (Bn ) = f −1 Bn ∈ f −1 (B),
n=1 n=1
(a) The forming of generated σ-algebras commutes with the forming of inverse images:
If E ⊆ P(Y ), then
σX f −1 (E) = f −1 σY (E) . (1.40)
(c) Let T be a topology on X and let B denote the corresponding Borel σ-algebra on
X. If B ⊆ X, then B|B is the Borel σ-algebra on B with respect to the relative
topology TB = T |B on B, i.e. B|B = σB (TB ).
Proof. (a): Since E ⊆ σY (E), one has f −1 (E) ⊆ f −1 (σY (E)), which implies the ⊆-part
of (1.40), as f −1 σY (E) is a σ-algebra by Prop. 1.55(b). To prove the ⊇-part of (1.40),
define n o
B := B ⊆ Y : f −1 (B) ∈ σX f −1 (E) .
Then E ⊆ B, B is a σ-algebra by Prop. 1.55(a),
implying σY (E) ⊆ B. This, in turn,
−1 −1 −1
yields f σY (E) ⊆ f (B) ⊆ σX f (E) , which is precisely the ⊇-part of (1.40),
completing the proof.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 41
(b): One merely has to apply (a) to the canonical inclusion map f : B −→ X, f (a) = a:
(1.39) (1.40) (1.39)
A|B = f −1 (A) = f −1 σX (E) = σB f −1 (E) = σB (E|B),
Please note the analogy between the restriction of a σ-algebra to a subset and the
forming of the relative topology on a subset.
Proposition 1.57. If (X, A, µ) is a measure space and A ∈ A, then the restriction
ν := µ↾A|A is a measure on (A, A|A).
Proof. First note that the statements make sense, since all intervals and all countable
subsets of Rn are Borel-measurable.
(a): For each k ∈ N, let ak := (a1 − k1 , . . . , an − k1 ), Ik :=]ak , b] ∈ I n . Then Ik ↓ I and
one computes
Yn Y n
n (1.10) n Ex. 1.20(b) 1
λ (I) = lim λ (Ik ) = lim b j − aj + = (bj − aj ),
k→∞ k→∞
j=1
k j=1
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 42
proving (a).
(b): If a ∈ Rn , then λn ({a}) = λn ([a, a]) = 0 by (a). Then (b) follows due to the
countable additivity of λn .
Thus, λn and β n satisfy Def. 1.1(a)(c). That they also satisfy Def. 1.1(b) will be seen
in Sec. 1.6.3.
1.5.2 Approximation
The Lebesgue measure λn has the benign property of being regular in the sense that,
with respect to λn , each A ∈ Ln can be approximated by open sets from the outside
and by closed and even by compact sets from the inside. Moreover, measurability of
A ⊆ Rn can be characterized by means of open and closed sets:
(a) For each A ∈ Ln and each ǫ ∈ R+ , there exist O ⊆ Rn open and F ⊆ Rn closed
such that
F ⊆ A ⊆ O ∧ λn (O \ A) < ǫ ∧ λn (A \ F ) < ǫ.
Thus, for each A ∈ Ln ,
(b) A set A ⊆ Rn is λn -measurable if, and only if, for each ǫ ∈ R+ , there exist O ⊆ Rn
open and F ⊆ Rn closed such that
F ⊆A⊆O ∧ λn (O \ F ) < ǫ.
∞
X ∞
X ǫ ǫ
n n
λ (O \ A) = λ (O) − λ (A) ≤ n n
λ (Ik ) + ǫ 2−k−1 − λn (A) < + = ǫ.
k=1 k=1
2 2
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 43
If λn (A) = ∞, then, for each k ∈ N, set Ak := A ∩ [−k, k]n . Using what we have
already shown, for each k ∈ N, thereSexists an open Ok ⊆ Rn such that Ak ⊆ Ok and
λn (Ok \ Ak ) < ǫ · 2−k . Letting O := ∞
k=1 Ok , we obtain O open with A ⊆ O and
∞
X ∞
X ∞
X
λn (O \ A) ≤ λn (Ok \ A) ≤ λn (Ok \ Ak ) < ǫ 2−k = ǫ,
k=1 k=1 k=1
as desired. One now obtains the closed set F by taking complements: To Ac , there
exists an open set U ⊂ Rn such that Ac ⊆ U and λn (U \ Ac ) < ǫ. Then F := U c is
closed, F ⊆ A, and λn (A \ F ) = λn (A ∩ U ) = λn (U \ Ac ) < ǫ, as needed. The first two
equalities of (1.41) are a direct consequence of what we have proved above. It remains
to show the third equality of (1.41). To this end, let 0 < α < λn (A). Then there exists
a closed F ⊆ A such that λn (F ) > α. For each k ∈ N, the set Ck := F ∩ [−k, k]n is
compact and Ck ↑ F . Thus limk→∞ λn (Ck ) = λn (F ) > α, i.e. λn (Ck ) > α must hold for
almost all of the Ck .
(b): First, assume A ∈ Ln . Given ǫ > 0, choose O and F as in (a), except with ǫ
replaced by ǫ/2. Then
ǫ ǫ
λn (O \ F ) = λn (O \ A) + λn (A \ F ) < + = ǫ
2 2
proves the first implication. For the converse, given A ⊆ Rn , for each k ∈ N, choose
n 1
a closed
S∞Fk and ann open OT k such that Fk ⊆ A ⊆ Ok and λ (Ok \ Fk ) < k . Then
∞ n n
B := k=1 Fk ∈ B , C := k=1 Ok ∈ B , B ⊆ A ⊆ C, and λ (C \ B) = 0. Thus,
A = B ∪ (A \ B) is the union of the Borel set B and the set A \ B, which is a subset of
the β n -null set C \ B. As λn is the completion of β n , A ∈ Ln .
According to Prop. 1.59(b), λn (A) = 0 for each countable A ⊆ Rn . For n > 1, Prop.
1.59(a) immediately provides examples of uncountable sets A ⊆ Rn with λn (A) = 0.
Finding uncountable subsets of R that are λ1 -null sets is not as simple. The Cantor set
C is an example. The Cantor set occurs frequently in the literature due to its many
interesting (e.g. topological) properties. The Cantor set C is what is left of the unit
interval [0, 1] after removing, in the initial step, the open middle third ] 31 , 32 [ and, then,
in each successive step, the open middle third of each remaining interval (after the first
step, one has left [0, 31 ]∪[ 23 , 1], after the second step, one has left [0, 91 ]∪[ 92 , 39 ]∪[ 96 , 97 ]∪[ 89 , 99 ],
etc.). We now give the precise definition:
Definition and Remark 1.61. Inductively, we define two sequences (In )n∈N0 and
(Kn )n∈N0 of subsets of [0, 1] such that, for each n ∈ N0 , the following statements hold
true:
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 44
[
˙ 2
n+1
[
˙ n [
˙ 2
m
Initially, we set I0,1 :=] 13 , 23 [, K0,1 := [0, 31 ], K0,2 := [ 32 , 1], clearly satisfying the above
conditions for n = 0. Now let n ∈ N0 and assume I0 , . . . , In as well as K0 , . . . , Kn to be
already defined such that (i) – (v) hold. Define, for each i ∈ {1, . . . , 2n+1 },
Then (ii) and (iii) for n + 1 are immediately clear as well as (v) for n + 1. Together
with (iv) for n, this implies (iv) for n + 1. Finally, (v) for n + 1 together with (ii) for n
implies (i) for n + 1, completing the inductive definition of (In )n∈N0 and (Kn )n∈N0 . We
are now in a position to define the Cantor set as
∞ 2[
\
n+1 ∞
\ [ [ m [ [ m
(iv) ˙ n ˙ 2 ˙ ∞ ˙ 2
C := Kn,i = [0, 1] \ Im,i = [0, 1] \ Im,i . (1.42)
m=0 i=1 m=0 i=1
n=0 i=1 n=0
Proposition 1.62. Let C ⊆ [0, 1] be the Cantor set as defined in (1.42). Then the
following holds:
(a) C is compact.
(b) C does not contain any nontrivial intervals, i.e. no intervals of positive length.
(c) λ1 (C) = 0.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 45
proving (c).
It is not difficult to see directly that all the endpoints αn,i and βn,i of the Kn,i are
elements of C, showing C must be infinite. However, there are only countably many
αn,i and βn,i , and, thus, to see that C is uncountable (Th. 1.64(b) below), we first have
to learn a bit more about the elements of C.
Lemma 1.63. Using the notation of Def. and Rem. 1.61, the left endpoints αn,i of Kn,i ,
n ∈ N0 , i ∈ {1, . . . , 2n+1 }, are precisely the numbers x ∈ R that have a finite triadic
expansion, where all digits are either 0 or 2: Let n ∈ N0 , x ∈ R. Then
n+1
!
X
∃ n+1 x = αn,i ⇔ ∃ x= dk · 3−k .
i∈{1,...,2 } d1 ,...,dn+1 ∈{0,2}
k=1
Similarly, the right endpoints βn,i of Kn,i , n ∈ N0 , i ∈ {1, . . . , 2n+1 }, are precisely the
numbers x ∈ R that have a periodic triadic expansion, where the period is 2 and all
digits are either 0 or 2: Let n ∈ N0 , x ∈ R. Then
n+1 ∞
!
X X
∃ n+1 x = βn,i ⇔ ∃ x= dk · 3−k + 2 · 3−k .
i∈{1,...,2 } d1 ,...,dn+1 ∈{0,2}
k=1 k=n+2
P
Proof. Since βn,i = αn,i + 3−n−1 and ∞ k=n+2 2 · 3
−k
= 2 · 3−n−2 · 32 = 3−n−1 , it suffices
to prove the statement for the αn,i . The statement for the αn,i is proved via induction
on n ∈ N0 : For n = 0, x = α0,1 if, and only if, x = 0 = 0 · 3−1 ; x = α0,2 if, and only if,
x = 32 = 2 · 3−1 . Thus, the statement holds for n = 0. For the induction step, assume
the statement to hold for some fixed n ∈ N0 . Then x = αn+1,2i−1 i ∈ {1, . . . , 2n+1 }, is
P, n+1
equivalent to x = αn,i and, by induction, this is equivalent to x = k=1 dk ·3−k +0·3−(n+2)
, 2n+1 }, is equivalent to
with d1 , . . . , dn+1 ∈ {0, 2}. Similarly, x = αn+1,2i , i ∈ {1, . . .P
x = αn,i + 2 · 3−n−2 and, by induction, this is equivalent to x = n+1 k=1 dk · 3
−k
+ 2 · 3−(n+2)
with d1 , . . . , dn+1 ∈ {0, 2}, completing the induction and the proof of the lemma.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 46
Theorem 1.64. Let C be the Cantor set defined in Def. and Rem. 1.61.
(a) C consists of all x ∈ [0, 1] that have a triadic expansion such that all digits are
either 0 or 2.
(b) C has the same cardinality as R; in particular, C is uncountable.
Proof. (a): Let x ∈ [0, 1]. According to the definition of C in (1.42), we have to show
n+1
2[
! ∞
!
X
−k
∀ x∈ Kn,i ⇔ ∃ x= dk · 3 .
n∈N0 (dk )k∈N in {0,2}
i=1 k=1
P∞−k
“⇐”: Suppose x = k=1 dk · 3 with dk ∈ {0, 2} for each k ∈ N. P Fix n ∈ N0 .
According to Lem. 1.63,Pthere exists i ∈ {1, . . . , 2n+1 } such that αn,i = n+1 −k
k=1 dk · 3 .
∞
Then αn,i ≤ x ≤ αn,i + k=n+2 2 · 3−k = βn,i , showing x ∈ Kn,i .
S n+1
“⇒”: Suppose x ∈ 2i=1 Kn,i for each n ∈ N0 . If x is a right endpoint βn,i , then x has
a triadic expansion of the required form by Lem. 1.63. If x is not a right endpoint βn,i ,
then, for each n ∈ N0 , there exists i ∈ {1, . . . , 2n+1 } such that αn,i ≤ x < αn,i + 3−n−1 .
Thus, x − αn,i < 3−n−1 , i.e. the coefficients of 3−1 , . . . , 3−(n+1) in thePtriadic expansions
of αn,i and x must be identical. Then Lem. 1.63 yields that x = ∞ k=1 dk · 3
−k
with
dk ∈ {0, 2} for each k ∈ N.
P
(b): Consider the map φ : {0, 2}N −→ C, φ((dk )k∈N ) := ∞ −k
k=1 dk · 3 . According to
(a), φ is well-defined and surjective. It is also injective by [Phi16a, Th. 7.99], proving
#{0, 2}N = #C. Thus
[Phi16a, Th. F.2]
#C = #{0, 2}N = #{0, 1}N = #R,
completing the proof of (b).
As mentioned at the beginning of this section, for n > 1, Prop. 1.59(a) provides examples
of sets A ⊆ Rn with #A = #R and λn (A) = 0. We have now also seen that the Cantor
set is a set A ⊆ R with #A = #R and λ1 (A) = 0. Since each λn is complete, this
implies #Ln = #P(R) for each n ∈ N. As #B n = #R according to [Els07, Sec. 2.8.3],
this also shows B n ( Ln .
Measurable maps are for measure theory what continuous maps are for topology: Recall
that a map f : X −→ Y between topological spaces (X, S) and (Y, T ) is continuous if,
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 47
and only if, f −1 (B) ∈ S for each B ∈ T . If one replaces each topology with a σ-algebra,
then one obtains the concept of a measurable map:
(a) f is called A-B-measurable (or often merely measurable) if, and only if, f −1 (B) ∈ A
for each B ∈ B.
(b) If S and T are topologies on X and Y , respectively, then f is called Borel measurable
if, and only if, it is σ(S)-σ(T )-measurable.
Proof. Since E ⊆ B, A-B-measurability implies (1.43). To prove the converse, note that
f −1 (E) ⊆ A implies
Th. 1.56(a)
f −1 (B) = f −1 σY (E) = σX f −1 (E) ⊆ A,
Example 1.67. (a) Constant maps are always measurable. More precisely, if (X, A)
and (Y, B) are measurable spaces, f : X −→ Y , f ≡ c, c ∈ Y , then, for each
B ∈ B, there are precisely two possibilities: c ∈ B, implying f −1 (B) = X ∈ A;
c∈/ B, implying f −1 (B) = ∅ ∈ A. Thus, f is A-B-measurable.
(b) Continuous maps are always Borel measurable: Let (X, S) and (Y, T ) be topological
spaces with corresponding Borel σ-algebras σ(S) and σ(T ). If f : X −→ Y is
continuous, then f −1 (U ) ∈ S ⊆ σ(S) for each U ∈ T and Prop. 1.66 implies f is
σ(S)-σ(T ) measurable. In particular, if M ⊆ Rn and f : M −→ Rm is continuous
(n, m ∈ N), then f is Borel measurable, i.e. B n -B m -measurable (cf. Ex. 1.58 and
Th. 1.56(c)).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 48
Theorem 1.69. Let (X, A, µ) be a measure space, let (Y, B) be a measurable space, and
let f : X −→ Y be A-B-measurable.
(b) The forming of push forward measures commutes with the composition of maps: If
(Z, C) is another measurable space, where g : Y −→ Z is B-C-measurable, then
Proof. (a): One has µf (∅) = µ(∅) = 0 and, if (Bn )n∈N is a sequence of disjoint sets in
B, then (f −1 (Bn ))n∈N is a sequence of disjoint sets in A, implying
∞
! ∞
!! ∞
! ∞
[ [ [ X
−1 −1
µf Bn = µ f Bn =µ f (Bn ) = µ f −1 (Bn )
n=1 n=1 n=1 n=1
∞
X
= µf (Bn ),
n=1
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 49
we have to verify µg◦f (C) = (µf )g (C) for each C ∈ C. To this end, for each C ∈ C, one
calculates
µg◦f (C) = µ (g ◦ f )−1 (C) = µ f −1 g −1 (C) = µf g −1 (C) = (µf )g (C),
Proposition 1.70. Let (X, A, µ) and (Y, B, ν) be measure spaces with completions
(X, Ã, µ̃) and (Y, B̃, ν̃). Moreover, let f : X −→ Y be A-B-measurable.
Proof. (a): Let B̃ ∈ B̃. Then there exist B, C ∈ B and N ⊆ C such that ν(C) = 0 and
B̃ = B ∪ N . Then
à := f −1 (B̃) = f −1 (B) ∪ f −1 (N ),
where f −1 (B) ∈ A and f −1 (N ) ⊆ f −1 (C) with µ(f −1 (C)) = 0, showing à ∈ Ã, i.e. f is
Ã-B̃-measurable.
(b): If ν = f (µ), and C ∈ B is a ν-null set, then µ(f −1 (C)) = ν(C) = 0. Thus, (a)
applies such that f is Ã-B̃-measurable and f (µ̃) is well-defined. If B̃, B, C, N are as in
the proof of (a), then
f (µ̃) (B̃) = µ̃ f −1 (B) ∪ f −1 (N ) = µ f −1 (B) = ν(B) = ν̃(B̃),
In the present section, we will show that the measures β n and λn satisfy Def. 1.1(b), i.e.
they are translation invariant. Indeed, in Th. 1.73 below, we will obtain the change of
β n and λn under arbitrary bijective affine maps, which will include the invariance of β n
and λn under Euclidean isometries, in particular translations, as a special case (also see
the paragraph before the statement of Th. 1.73 for a brief review of the relation between
affine maps, Euclidean isometries, and translations). For A ∈ P(Rn ), x ∈ Rn , n ∈ N,
we recall the notation
A + x = {a + x : a ∈ A}.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 50
(a) β n (resp. λn ) is translation invariant and, moreover, it is the only translation in-
variant measure µ on B n (resp. Ln ) satisfying µ([0, 1]n ) = 1.
(b) If µ is a translation invariant measure on B n (resp. Ln ) such that µ([0, 1]n ) = α <
∞, then µ = αβ n (resp. µ = αλn ).
Thus, the σ-finite measures β n and T−a (β n ) agree on the semiring I n . By Cor. 1.46(a),
they must also agree on B n , proving the translation invariance of β n . Then, according
to Prop. 1.70(a), Ta is Ln -Ln -measurable. Now apply Prop. 1.70(b) with µ := β n and
ν := T−a (β n ) = β n and obtain
T−a (λn ) = T−a (µ̃) = ν̃ = µ̃ = λn ,
proving the translation invariance of λn . Conversely, assume µ : B n −→ [0, ∞] to
be a translation invariant measure. In a first step, we assume µ(]0, 1]n ) = 1. Given
n n
n ) ∈ N , we can decompose ]0, 1] into n1 · · · nn translative copies of the
g := (n1 , . . . , nQ
n 1
interval Ig := i=1 ]0, ni ]: If
k1 kn n n
Kg := ,..., ∈ Q : (k1 , . . . , kn ) ∈ (N0 ) , ∀ 0 ≤ ki < ni ,
n1 nn i∈{1,...,n}
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 51
Thus, the σ-finite measures µ and β n agree on IQn , i.e. they must agree on B n by Cor.
1.46(a). If µ extends to a measure on Ln , then µ = λn by Th. 1.51. It remains to
consider a translation invariant measure µ on B n (resp. on Ln ) with µ([0, 1]n ) = 1. Set
α := µ(]0, 1]n ). As µ is translation invariant, we obtain
showing µ ≡ 0 ≡ 0 · β n (resp. µ ≡ 0 ≡ 0 · λn ).
Note that Th. 1.72(b) does not hold without the condition α < ∞, as counting measure
is a translation invariant measure on both B n and Ln that is not a multiple of β n (resp.
λn ).
In generalization of the above result on the translation invariance of β n and λn , we will
now investigate these measures’ behavior under general bijective affine maps. Recall
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 52
from Linear Algebra (cf. [Phi19b, Sec. 1.5]) that a map f : Rn −→ Rn is called affine
if, and only if, there exists a linear map L : Rn −→ Rn and a vector a ∈ Rn such that
f (x) = L(x) + a for each x ∈ Rn . Thus, using Def. 1.71(a), f : Rn −→ Rn is affine if,
and only if, f = Ta ◦ L with a linear L and a vector a as above. Clearly, in that case, a
and L are uniquely determined by the affine f (a = f (0), L = f − a), and one defines
det f := det L. In consequence, one has the equivalences
f bijective ⇔ L bijective ⇔ det L 6= 0 ⇔ det f 6= 0.
Theorem 1.73. Let n ∈ N, f : Rn −→ Rn .
Proof. Since we know from Linear Algebra that f is a Euclidean isometry if, and only
if, f = L + a with an orthogonal linear map L : Rn −→ Rn and a ∈ Rn (cf. [Phi19b,
Cor. 10.37]), where an orthogonal linear map satisfies | det L| = 1 (cf. [Phi19b, Prop.
10.31(b)]), it suffices to prove (a).
Since the affine map f is continuous, it is B n -B n -measurable. As f is affine and bijective,
there exists a linear map L : Rn −→ Rn with det L 6= 0 and some a ∈ Rn such that
f = Ta ◦ L. The main work consists of showing
L(β n ) = | det L|−1 β n . (1.49)
Let µ : B n −→ [0, ∞] be some arbitrary translation invariant measure on B n . We claim
L(µ) to be translation invariant as well: Indeed, for each A ∈ B n and each v ∈ Rn , using
the linearity of L and the translation invariance of µ,
L(µ) (A + v) = µ L−1 (A + v) = µ L−1 (A) + L−1 (v) = µ L−1 (A) = L(µ) (A).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 53
We introduce the abbreviation I := [0, 1]n and, for the rest of the proof, we make the
additional assumption that αµ := µ(I) < ∞. Then µ = αµ β n by Th. 1.72(b), and
αL,µ := L(µ) (I) = µ L−1 (I) = αµ β n L−1 (I) < ∞
(as the continuous image L−1 (I) of the compact set I is compact), implying
L(µ) = αL,µ β n .
We need to show αL,β n = | det L|−1 . Suppose L is orthogonal (i.e. | det L| = 1) and let
B := {x ∈ Rn : kxk2 ≤ 1}. Then, since L is a Euclidean isometry, L−1 (B) = B and,
thus,
αL,µ β n (B) = L(µ) (B) = µ L−1 (B) = µ(B) = αµ β n (B),
implying αL,µ = αµ , thereby proving (1.49) for orthogonal L. Next, assume L is repre-
sented by a diagonal matrix with diagonal entries d1 , . . . , dn ∈ R+ with respect to the
standard basis (e1 , . . . , en ) of Rn . Then
n
Y
−1
α L,β n
n n
= β L (I) = β [(0, . . . , 0), (d−1 −1
1 , . . . , dn )] = d−1 −1
j = | det L| ,
j=1
proving (1.49) also for this case. Now let L be a general linear map with det L 6= 0. We
obtain this general case from the above special cases, using a bit more Linear Algebra:
Suppose the matrix M ∈ M(n, R) represents L with respect to the standard basis of
Rn . According to [Phi19b, Th. 11.7], P := M M t is symmetric and positive definite
(in particular, all eigenvalues of P are in R+ by [Phi19b, Th. 11.3(a)]). Using [Phi19b,
Th. 10.41], there exists an orthogonal matrix U ∈ On (R) such that P can be written
as P = U DU t , where D is a diagonal matrix with diagonal entries d1 , . . . , dn ∈ R+ as
above. With respect to the standard basis of Rn , P , U , and D represent linear maps
LP , LU , and LD , respectively, where LU is orthogonal (cf. [Phi19b, Prop. 10.36]).
√ If√S is
the linear map represented by the diagonal matrix with diagonal entries d1 , . . . , dn ,
then S 2 = LD and K := S −1 (LU )t L is orthogonal:
P L
z}|{
−1
KK = St
(LU ) LLt LU S −1 = S −1 (LU )t LU S 2 (LU )t LU S −1 = Id .
t
| {z } | {z }
Id Id
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 54
completing the proof of (1.49). We now come back to our bijective affine map f =
Ta ◦ L. Since L(β n ) is translation invariant by (1.49), f (βn ) = Ta (L(β n )) = L(β n ) =
| det f |−1 β n as claimed. Then Prop. 1.70(b) yield with µ := β n and ν := f (β n ) =
| det f |−1 β n and obtain
(1.37c)
f (λn ) = f (µ̃) = ν̃ = | det f |−1 µ̃ = | det f |−1 λn ,
completing the proof of (1.47a). Finally, (1.47b) follows by applying (1.47a) to f −1 :
∀ n λn f (A) = f −1 (λn ) (A) = | det f | λn (A),
A∈L
We are now in a position to show the existence of subsets M of Rn that are not λn -
measurable and, more generally, that no ideal n-dimensional volume as defined in Def.
1.1 can exist (see Th. 1.74 below). The proof hinges on the validity of the axiom of
choice AC. Without the axiom of choice the situation becomes subtle and difficult: It
has been shown that there exist models of ZF set theory (without AC), where every
subset of Rn is λn -measurable. However, these constructions assume the existence of
so-called inaccessible cardinals, and if the existence of such cardinals is consistent with
ZF remains unknown (see the comments and references at the end of [Els07, Sec. III.3]).
In preparation for Th. 1.74, let G ⊆ Rn , n ∈ N, be an additive subgroup of Rn and
recall from Linear Algebra (cf. [Phi19a, Def. 4.26]) that the factor group (or quotient
group) Rn /G consists of all cosets
G + a = {g + a : g ∈ G}, a ∈ Rn .
These are precisely the equivalence classes of the equivalence relation defined by x ∼
y :⇔ x − y ∈ G. By AC, there exists a set MG ⊆ Rn of representatives of Rn /G,
containing precisely one element from each coset G + a ∈ Rn /G.
Theorem 1.74. Let n ∈ N.
(a) Let (Rn , A, µ) be a measure space such that µ extends β n (i.e. B n ⊆ A and µ↾Bn =
β n ). Moreover, let G := Qn ⊆ Rn . We consider G as an additive subgroup of Rn
and let M := MG be a set of representatives of Rn /G. If µ is translation invariant
with respect to G (i.e. A + g ∈ A and µ(A + g) = µ(A) hold for each A ∈ A and
each g ∈ G), then M ∈ / A and µ(A) = 0 for each A ∈ A with A ⊆ M (in particular,
choosing A := L , µ := λn , yields the existence of a set M ⊆ Rn (by use of AC)
n
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 55
(b) There does not exist an ideal n-dimensional volume as defined in Def. 1.1.
(c) Let η n : P(Rn ) −→ [0, ∞] denote the outer measure induced by λn . Then each
A ⊆ Rn with η n (A) > 0 contains a set B that is not Lebesgue measurable (i.e.
B∈ / Ln ).
implying µ(M0 ) > 0. On the other hand, since G ∩ I is infinite (and countable),
!
X X [
µ(M0 ) = µ(g + M0 ) = µ (g + M0 )
g∈G∩I g∈G∩I g∈G∩I
n
≤ µ ]0, 2] = β n ]0, 2] n
= 2n < ∞,
Since, for each g ∈ G, g + M is still a set of representatives of Rn /G, (a) implies
η n A ∩ (g + M ) = λn A ∩ (g + M ) = 0 if A ∩ (g + M ) ∈ Ln . Thus, if A ∩ (g + M ) ∈ Ln
for each g ∈ G, then (1.50) implies η n (A) = 0.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 56
Recall that, in [Phi16b, Ex. 1.52(b)], we extended the total order on R to a total order
on R = R ∪ {−∞, ∞}. Then the resulting order topology T on R also extended the
standard topology on R, where a base for T is given by the set of open intervals
I o := R ∪ ]a, b[: a, b ∈ R, a < b
∪ [−∞, a[: a ∈ R ∪ {∞} ∪ ]a, ∞] : a ∈ R ∪ {−∞} (1.51)
Notation 1.75. Let B denote the Borel sets on R, i.e. the Borel sets with respect to
the order topology T on R.
Lemma 1.76. For the Borel sets on R, we have the following identities:
B = B ∪ E : B ∈ B 1 , E ⊆ {−∞, ∞} , (1.52a)
B = σ({[−∞, α[: α ∈ R}) = σ({[−∞, α] : α ∈ R})
= σ({]α, ∞] : α ∈ R}) = σ({[α, ∞] : α ∈ R}), (1.52b)
B|R = B 1 . (1.52c)
Proof. (1.52a): Let A denote the right-hand side of (1.52a). Since {−∞} and {∞} are
closed sets in R, every E ⊆ {−∞, ∞} is a Borel set, implying A ⊆ B. To verify the
remaining inclusion, note σR (I o ) = B, I o ⊆ A, and A is a σ-algebra.
(1.52b): Clearly, all four generator sets are subsets of B. Let E1 := {[−∞, α[: α ∈ R},
E2 := {[−∞, α] : α ∈ R}. We first show σ(E2 ) = B: If α, β ∈ R, then ]α, β] =
[−∞, β] \ [−∞, α]. Thus, I 1 ⊆ σ(ET2 ), implying B 1 ⊆ σ(E2 ). Since also {−∞} =
T
n∈N [−∞, −n] ∈ σ(E2 ) and {∞} = S n∈N
]n, ∞] ∈ σ(E2 ), σ(E2 ) = B now follows from
(1.52a). If α ∈ R, then [−∞, α[= n∈N [−∞, α − n1 ] ∈ σ(E1 ) implies σ(E1 ) = B. Since
the elements of {]α, ∞] : α ∈ R} are precisely the complements of the elements of E2
and the elements of {[α, ∞] : α ∈ R} are precisely the complements of the elements of
E1 , the proof of (1.52b) is complete.
(1.52c): Since the topology on R is the relative topology inherited from R, (1.52c) follows
from Th. 1.56(c). Alternatively, (1.52c) is immediate from (1.52a).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 57
(b) A function f : X −→ C is called measurable if, and only if, it is A-B 2 -measurable,
where B 2 are the Borel sets on C with respect to the standard topology on C (the
notation B 2 is reasonable, since C = R2 and B 2 is precisely the set of Borel sets on
R2 ).
Notation 1.78. In the present context, it is useful to introduce a shorthand notation for
certain subsets of X: Let F := (fi )i∈I be a family of functions defined on X. Let P (F)
denote a sentence (i.e. a sequence of symbols) and let P (F)(x) denote the corresponding
sentence, where each occurrence of fi in P (F) was replaced by fi (x). If P (F) is such
that P (F)(x) constitutes a statement (i.e. is either true or false) for each x ∈ X, then
we define
{P (F)} := {x ∈ X : P (F)(x) holds true}. (1.53)
For example, consider f, g : X −→ R. If F = (f ) and α ∈ R, then, e.g.,
Theorem 1.79. Let (X, A) be a measurable space and f : X −→ R. Then the following
statements are equivalent:
(i) f is measurable.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 58
Notation 1.80. Let X be a set and let (fi )i∈I be a family of functions fi : X −→ R.
Then define
inf fi : X −→ R, inf fi (x) := inf{fi (x) : i ∈ I}, (1.55a)
i∈I i∈I
sup fi : X −→ R, sup fi (x) := sup{fi (x) : i ∈ I}. (1.55b)
i∈I i∈I
where (1.55e) is defined, provided each limit limi→∞ fi (x) exists in R (of course, this is
the pointwise limit of the fi , which we have already studied previously). The following
notation will also turn out to be useful: If f : X −→ R, then define
fi ↑ f :⇔ f = lim fi ∧ ∀ f1 (x) ≤ f2 (x) ≤ . . . , (1.56a)
i→∞ x∈X
fi ↓ f :⇔ f = lim fi ∧ ∀ f1 (x) ≥ f2 (x) ≥ . . . . (1.56b)
i→∞ x∈X
In the case of (1.55e) (and, in particular, (1.56)), we call the convergence uniform if,
and only if,
∀+ ∃ ∀ ∀ fi (x) − f (x) < ǫ. (1.57)
ǫ∈R N ∈N i≥N x∈X
Lemma 1.81. Let X be a set and let (fk )k∈N be a sequence of functions fk : X −→ R.
Then
\ ∞ \ ∞
!
∀ inf fk ≥ α = {fk ≥ α} ∧ sup fk ≤ α = {fk ≤ α} , (1.58a)
α∈R k∈N k∈N
k=1 k=1
lim fk = sup inf fi ∧ lim fk = inf sup fi . (1.58b)
k→∞ k∈N i≥k k→∞ k∈N i≥k
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 59
Proof. Let α ∈ R, x ∈ X. Then the two assertions in (1.58a) are precisely the equiva-
lences
For each k ∈ N, let rk := inf{fi (x) : i ≥ k}. Then α ≥ rk : If there exists i0 ≥ k such
that α ≥ fi0 (x), then α ≥ fi0 (x) ≥ rk , since rk is a lower bound for the fi (x), i ≥ k.
If α < fi (x) for all i ≥ k, then α < ∞ and α = rk = inf{fi (x) : i ≥ k}, since, for
α > −∞, each ]α, α + ǫ[, ǫ > 0, must contain some fi (x), i ≥ k, and, for α = −∞, each
] − ∞, −n[, n ∈ N, must contain some fi (x) (as α is a cluster point). Now α ≥ rk for
each k ∈ N implies β ≤ α, since β is the smallest upper bound of the rk . It remains to
show α ≤ β. To this end, let γ ∈ R be an arbitrary upper bound of R := {rk : k ∈ N}.
As β is an upper bound of R, it suffices to show α ≤ γ. Seeking a contradiction, assume
γ < α. For each k ∈ N, since rk ≤ γ and rk = inf{fi (x) : i ≥ k}, there exists i ≥ k such
that fi (x) < β, showing that [α, rk [ must contain fj (x) for infinitely many j ∈ N, in
contradiction to α being the smallest cluster point of (fj (x))j∈N . Thus, we have shown
α = β, as desired. One can conclude the second equation of (1.58b) from the first by
applying it to the reversed order on R. Alternatively, one can conduct an analogous
proof as for the first equation: Let x ∈ X. We have to show
n o
α := lim sup fk (x) = inf sup{fi (x) : i ≥ k} : k ∈ N =: β.
k→∞
For each k ∈ N, let rk := sup{fi (x) : i ≥ k}. Then α ≤ rk : If there exists i0 ≥ k such
that α ≤ fi0 (x), then α ≤ fi0 (x) ≤ rk , since rk is an upper bound for the fi (x), i ≥ k.
If α > fi (x) for all i ≥ k, then α > −∞ and α = rk = sup{fi (x) : i ≥ k}, since, for
α < ∞, each ]α − ǫ, α[, ǫ > 0, must contain some fi (x), i ≥ k, and, for α = ∞, each
]n, ∞[, n ∈ N, must contain some fi (x) (as α is a cluster point). Now α ≤ rk for each
k ∈ N implies β ≥ α, since β is the largest lower bound of the rk . It remains to show
α ≥ β. To this end, let γ ∈ R be an arbitrary lower bound of R = {rk : k ∈ N}. As
β is a lower bound of R, it suffices to show α ≥ γ. Seeking a contradiction, assume
γ > α. For each k ∈ N, since rk ≥ γ and rk = sup{fi (x) : i ≥ k}, there exists i ≥ k
such that fi (x) > β, showing that ]rk , α] must contain fj (x) for infinitely many j ∈ N,
in contradiction to α being the largest cluster point of (fj (x))j∈N . Thus, we have shown
α = β, as desired, finishing the proof of (1.58b). If the limit of the sequence (fk (x))k∈N ,
x ∈ X, exists, then it is the only cluster point of the sequence, proving (1.58c).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 60
Theorem 1.82. Let (X, A) be a measurable space, let (fk )k∈N be a sequence of measur-
able functions, fk : X −→ R.
(a) Each of the functions inf k∈N fk , supk∈N fk , limk→∞ fk , limk→∞ fk is measurable. If
f = limk→∞ fk , f : X −→ R, then f is measurable as well.
Proof. (a): If fk is measurable, then {fk ≥ α} ∈ A for each α ∈ R. Thus, inf k∈N fk
is measurable by (1.58a) and Th. 1.79(v); supk∈N fk is measurable by (1.58a) and Th.
1.79(iii). In consequence, limk→∞ fk and limk→∞ fk are measurable by (1.58b), implying
limk→∞ fk to be measurable by (1.58c).
(b) follows by applying the inf and sup parts of (a) to the sequence of measurable
functions (f1 , . . . , fn , fn , fn , . . . ).
(a) A function f : X −→ Rn is measurable if, and only if, all coordinate functions
f1 , . . . , fn : X −→ R are measurable.
(b) A function f : X −→ C is measurable if, and only if, the real-valued functions Re f
and Im f are both measurable.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 61
{f = g} = {f ≤ g} ∩ {g ≤ f } ∈ A, {f 6= g} = {f = g}c ∈ A.
If α ∈ R, then {f < α} = {−f > −α}, showing −f to be measurable by Th. 1.79.
Now assume f, g to be R-valued. Then h := (f, g) : X −→ R2 is measurable by Th.
1.83(a). Since the continuous functions s, p : R2 −→ R, s(x, y) := x + y, p(x, y) := xy
are B 2 -B 1 -measurable, f + g = s ◦ h and f g = p ◦ h are measurable. If f, g are R-valued,
then we have
A0 := {f + g = 0} = {f = −g} ∈ A,
A1 := {f + g = ∞} = {f = ∞} ∩ {g 6= −∞} ∪ {g = ∞} ∩ {f 6= −∞} ∈ A,
A2 := {f + g = −∞} = {f = −∞} ∩ {g 6= ∞} ∪ {g = −∞} ∩ {f 6= ∞} ∈ A.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 62
(a) f : X −→ R is measurable if, and only if, both f + and f − are measurable.
Caveat 1.86. One has to use caution with the statement “The composition of measur-
able functions is measurable.”: While it is correct if it is meant in the sense of Prop.
1.68, it is false if applied to functions f, g : R −→ R, that are measurable in the sense
of Def. 1.77(d): If f and g are L1 -B 1 -measurable, then g ◦ f is not necessarily L1 -B 1 -
measurable (see, e.g., [RF10, Sec. 3.1, p. 57] for a counterexample). Of course, g ◦ f is
L1 -B 1 -measurable, if g is B 1 -B 1 -measurable (in particular, if g is continuous).
—
In the rest of the present section, we study so-called simple or step functions (Def. 1.87).
Their importance lies in the fact that every R-valued measurable map is the pointwise
limit of a suitable sequence of simple functions (see Th. 1.89 below). We will make use
of this result, when developing the integration theory of R-valued measurable maps (see
Sections 2.1.1 and 2.1.2).
Caveat: The functions in S and S + are R-valued, but the functions in M and M+ are
allowed to be R-valued.
—
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 63
(c) A function f : X −→ R is a simple function if, and only if, it is a linear combina-
tion of characteristic functions of measurable subsets of X.
For β1 , . . . , βn 6= 0, also
#(f /g)(X) = # αi /βj : i ∈ {1, . . . , m}, j ∈ {1, . . . , n} ≤ m · n < ∞.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 64
(a) f ∈ M+ if, and only if, there exists a sequence (φk )k∈N in S + with φk ↑ f .
(b) f ∈ M with f bounded from below by α ∈ R (and bounded from above by β ∈ R) if,
and only if, there exists a sequence (φk )k∈N in S with all φk ≥ α (and all φk ≤ β)
and φk ↑ f .
(c) f ∈ M if, and only if, there exists a sequence (φk )k∈N in S with f = limk→∞ φk .
Additionally, in each case, if f is also bounded, then one can obtain the φk to converge
uniformly to f .
Proof. Note that (a) is a special case of (b) (to obtain (a) from (b), choose α := 0 and,
in the bounded case, also choose β ≥ 0 to be some upper bound of f in (b)). We, thus,
proceed to prove (b): If (φk )k∈N is a sequence in S that converges pointwise to f , then,
as the simple functions φk are measurable, f is measurable by Th. 1.82(a). Moreover, if,
for x ∈ X, α ≤ φk (x) (resp. φk (x) ≤ β) for each k ∈ N, then, α ≤ f (x) (resp. f (x) ≤ β)
as well. For the converse, given f , we have to construct suitable simple functions φk .
To this end, we define the A-measurable sets
(
{α + 2jk ≤ f < α + j+1
2k
} for 0 ≤ j ≤ k · 2k − 1,
∀ ∀ Ak,j :=
k∈N j∈{0,1,...,k·2k } {f ≥ α + k} for j = k · 2k
(as f is assumed to be measurable, the Ak,j are A-measurable by Th. 1.79). Clearly, for
each fixed k ∈ N, the Ak,j form a decomposition of X:
[
˙ k·2
k
∀ X= Ak,j . (1.61)
k∈N j=0
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 65
Since X is the disjoint union of the Ak,j by (1.61), this immediately implies
∀ α ≤ φ k ≤ α + k · 2k .
k∈N
We next verify that the sequence (φk )k∈N is increasing: Note that, for each k ∈ N, we
have the disjoint unions
Ak,j = Ak+1,2j ∪˙ Ak+1,2j+1 for 0 ≤ j ≤ k · 2k − 1,
[
˙ (k+1)·2
k+1
Ak,k·2k = Ak+1,l .
l=k·2k+1
k
Since 2jk ≤ 2k+1
l
for l ∈ {2j, 2j + 1}, and k·2
2k
l
≤ 2k+1 for l ∈ {k · 2k+1 , . . . , (k + 1) · 2k+1 },
we have φk (x) ≤ φk+1 (x) for each x ∈ X. We now fix x ∈ X and show limk→∞ φk (x) =
f (x): For each k ∈ N, there exists a unique j = j(k, x) such that x ∈ Ak,j . Then
φk (x) = α + 2jk ≤ f (x), showing f (x) to be an upper bound for the φk (x). If f (x) = ∞,
then, for each k ∈ N, φk (x) = α + k → ∞. On the other hand, if f (x) < ∞, then, given
ǫ > 0, we can choose k0 ∈ N such that 2k10 < ǫ and f (x) < α + k0 . If k ≥ k0 , then
f (x) − φk (x) < 2−k ≤ 2−k0 < ǫ,
proving limk→∞ φk (x) = f (x). Under the additional hypothesis f ≤ β ∈ R, one can
choose α + k0 ≥ β, independently of x, showing uniform convergence. Moreover, in this
case, the above observation φk (x) ≤ f (x) for each k ∈ N and each x ∈ X also yields
φk (x) ≤ β for each k ∈ N, x ∈ X.
(c): As before, if f = limk→∞ φk with (φk )k∈N in S, then f ∈ M by Th. 1.82(a). For
the converse, we write f ∈ M as f = f + − f − and, by (a), choose (φk )k∈N in S + with
φk ↑ f + and (ψk )k∈N in S + with ψk ↑ f − (with uniform convergence in each case for f
bounded). Then f = limk→∞ (φk − ψk ) (with uniform convergence for f bounded).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 66
Proposition 1.91. In the situation of Def. 1.90 above, for each i ∈ I, let Ei be a
generator of Ai .
(b) If I is finite and, for each i ∈ I, there exists a sequence (Ei,k )k∈N in Ei such that
∞
[
Xi = Ei,k , (1.62)
k=1
Proof. (a): As Ep∗ ⊆ Ep , we have σ(Ep∗ ) ⊆ σ(Ep ). For the remaining inclusion, we note
that, by the definition
of Ep∗ , one
has πi−1 (Ei ) ⊆ Ep∗ for each i ∈ I. Moreover, by Th.
1.56(a), σ πi−1 (Ei ) = πi−1 σ(Ei ) = πi−1 (Ai ), implying σ(Ep ) ⊆ σ(Ep∗ ).
(b): From (a), we know Ap = σ(Ep∗ ). If i ∈ I and Ei ∈ Ei , then πi−1 (Ei ) ∈ Ep∗ by the
T
definition of Ep∗ . Thus, if Ei ∈ Ei for each i ∈ I, then i∈I πi−1 (Ei ) ∈ σ(Ep∗ ), since I is
finite. This already yields σ(Fp ) ⊆ σ(Ep∗ ) (for this inclusion, we actually did not use
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 67
(1.62)). For the remaining inclusion, we use the finiteness of I together with (1.62) to
write every element of Ep∗ as a countable union of elements from Fp :
[ Y
∀ ∀ πi−1 (Ei ) = Ei × Eα,kα ∈ σ(Fp ).
i∈I Ei ∈Ei ∪{Xi }
(kα )α∈I\{i} ∈NI\{i} α∈I\{i}
Example 1.92. Let p, q ∈ N. For the Borel sets on Rp and Rq , we obtain B p+q = B p ⊗B q
(also see Th. 1.94(c) below): We know from Ex. 1.8(c), that the Borel sets are generated
by the compact intervals: B p = σ(Icp ), B q = σ(Icq ), B p+q = σ(Icp+q ). Thus, we can apply
Prop. 1.91(b) with E1 = Icp , E2 = Icq , and Fp = Icp+q to obtain
Theorem 1.93. Let I be a nonempty index set and, for eachQi ∈ I, let (Xi , Ai ) be a
measurable space. We consider the Cartesian product X := i∈I Xi with the product
σ-algebra Ap = σ(Ep ) according to Def. 1.90.
(b) Ap is the smallest σ-algebra A on X such that all projections πi , i ∈ I, are A-Ai -
measurable.
(c) If (Y, B) is a measurable space and f : Y −→ X, then the following statements are
equivalent:
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
1 MEASURE THEORY 68
Theorem 1.94. Let I be a nonempty index set and, for each i ∈ I, let (Xi , Ti ) be a
Q space. Let Ai denote the corresponding Borel σ-algebra on Xi . Moreover, let
topological
X := i∈I Xi , let T denote the
Nproduct topology on X, let AT be the resulting Borel
σ-algebra on X, and let Ap := i∈I Ai be the product σ-algebra of the Ai .
(b) If I is countable and each (Xi , Ti ) is a C2 space (i.e. each Ti has a countable base),
then Ap = AT .
Proof. (a): Since each projection πi , i ∈ I, is continuous, it is πi−1 (A) ∈ T for each
A ∈ Ti , implying it to be Ai -AT -measurable. Then Ap ⊆ AT follows from Th. 1.93(b).
(b): Due to (a), it remains to show AT ⊆ Ap . For each i ∈ I, let Bi be a countable base
of Ti . According to [Phi16b, Ex. 1.53(a)], a base for T is given by
( )
\
B := πj−1 (Bj ) : J ⊆ I, 0 < #J < ∞, ∀ Bj ∈ Bj .
j∈J
j∈J
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 69
2 Integration
The definition of the Lebesgue integral for nonnegative simple functions makes use of
representations of the integrands, where Lem. 2.2 then states that the value of the
integral does actually not depend on the representation of the integrated function.
Definition 2.1. Let (X, A, µ) be a measure space. Moreover, let f : X −→ R+
0 be a
+
simple function (i.e. f ∈ S (A)), where
N
X
f= αi χAi (2.1)
i=1
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 70
Lemma 2.2. Let (X, A, µ) be a measure space. Then the value of the integral in
(2.2) does not depend on the representation of the simple function f : If M, N ∈ N;
α1 , . . . , αN ≥ 0; β1 , . . . , βM ≥ 0; A1 , . . . , AN ∈ A; and B1 , . . . , BM ∈ A are such that
N
X M
X
f= αi χAi = βi χBi , (2.3a)
i=1 i=1
then
N
X M
X
αi µ(Ai ) = βi µ(Bi ). (2.3b)
i=1 i=1
Note that two distinct elements of D are always disjoint: Indeed, if A, B ∈ D with
A 6= B, then there exists i0 ∈ {1, . . . , N + M } such that A ⊆ Ai0 and B ⊆ Aci0 . Next,
we consider i ∈ {1, . . . , N + M } and x ∈ Ai . Then, for each j ∈ {1, . . . , N + M }, either
x ∈ Aj or x ∈ Acj , implying
(N +M )
[ \
Ai = Mj : Mj = Ai ∧ ∀ Mj ∈ {Aj , Acj } .
j∈{1,...,N +M }
j=1
implying
K
X X X
f= γ i χCi , where γi = αj = βj
i=1 j∈{1,...,N }: j∈{1,...,M }:
Ci ⊆Aj Ci ⊆Bj
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 71
and
N
X N
X X K
X
αi µ(Ai ) = αi µ(Cj ) = γj µ(Cj )
i=1 i=1 j∈{1,...,K}: j=1
Cj ⊆Ai
M
X X M
X
= βi µ(Cj ) = βi µ(Bi ),
i=1 j∈{1,...,K}: i=1
Cj ⊆Bi
proving (2.3b).
Lemma 2.3. Let (X, A, µ) be a measure space.
R
(a) For each A ∈ A, one has A χA dµ = µ(A).
(b) For each f, g ∈ S + (A) and each α, β ∈ R+
0 , one has
Z Z Z
(αf + βg) dµ = α f dµ + β g dµ .
X X X
implying
Z Z N
X M
X Z
α f dµ + β g dµ = ααi µ(Ai ) + ββi µ(Bi ) = (αf + βg) dµ ,
X X i=1 i=1 X
proving (b).
(c): If f ≤ g, then g − f ∈ S + (A). Thus, (b) applies to g = f + (g − f ), yielding
Z Z Z Z
g dµ = f dµ + (g − f ) dµ ≥ f dµ ,
X X X X
as desired.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 72
Note that, so far, we have not used the σ-additivity of the measure µ – it would have
sufficed to take µ to be a content on an algebra. However, this will change in the
following section.
As for nonnegative simple functions above, the definition of the Lebesgue integral for
nonnegative measurable functions also makes use of representations of the integrand,
where, once again, a subsequent lemma (Lem. 2.5) then states that the value of the
integral does actually not depend on the representation of the integrated function.
(b) The value of the integral in (2.4) does not depend on the representation of the non-
negative measurable function f ∈ M+ (A): If (φi )i∈N and (ψi )i∈N both are sequences
in S + (A) with φi ↑ f and ψi ↑ f , then
Z Z
lim φi dµ = lim ψi dµ . (2.6)
i→∞ X i→∞ X
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 73
P
v= N +
i=1 αi χAi with αi ∈ R0 , Ai ∈ A, N ∈ N, and use µ being continuous from below
to compute
Z N
X N
X Z
v dµ = αi µ(Ai ) = lim αi µ(Ai ∩ Bn ) = lim v · χBn dµ
X n→∞ n→∞ X
i=1 i=1
Lem. 2.3(c)
Z Lem. 2.3(c)
Z
≤ lim β un · χBn dµ ≤ β lim un dµ . (2.7)
n→∞ X n→∞ X
As (2.7) holds for each β > 1, it must still hold for β = 1, proving (2.5).
(b): For each k ∈ N, we have φk ≤ f = limi→∞ ψi . Thus, by (a),
Z Z
∀ φk dµ ≤ lim ψi dµ ,
k∈N X i→∞ X
implying Z Z
lim φk dµ ≤ lim ψi dµ . (2.8)
k→∞ X i→∞ X
Interchanging the roles of φi and ψi above, yields (2.8) with the inequality reversed,
proving (2.6).
(b) For each f, g ∈ M+ (A) and each α, β ∈ [0, ∞], one has
Z Z Z
(αf + βg) dµ = α f dµ + β g dµ .
X X X
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 74
Proof. (a): Choose a sequence (φn )n∈N in S + (A) such that φn ↑ f . Set A := {f > 0}.
As f is measurable, we have A ∈ A. If µ(A) > 0, then, letting An := {f > n1 }, we have
An ↑ A and, thus, µ(An0 ) > 0 for some n0 ∈ N. In consequence,
Z Z Z
Lem. 2.5(a) 1 1
f dµ = lim φn dµ ≥ lim χAn dµ ≥ µ(An0 ) > 0.
X n→∞ X n→∞ X n n0
R R
showing ∞ · X f dµ = ∞R = X ∞ · f dµ . IfR µ(A) = 0, then µ({φn > 0}) ≤ µ(A) = 0
for each n ∈ N, implying X f dµ = limn→∞ X φn dµ = 0.
(b): Choose sequences (φi )i∈N and (ψi )i∈N in S + (A) such that φi ↑ f and ψi ↑ f . Then
(φi + ψi ) ↑ (f + g), implying
Z Z Z Z
Lem. 2.3(b)
(f + g) dµ = lim (φi + ψi ) dµ = lim φi dµ + lim ψi dµ
X i→∞ X i→∞ X i→∞ X
Z Z
= f dµ + g dµ .
X X
Similarly, if α ∈ R+ +
0 , then αφi ∈ S (A) for each i ∈ N, (αφi ) ↑ (αf ), and
Z Z Z Z
Lem. 2.3(b)
α f dµ = lim α φi dµ = α lim φi dµ = α f dµ .
X i→∞ X i→∞ X X
It remains to consider the case α = ∞. As in the proof of (a), set A := {f > 0}. Then
n · χA ↑ ∞ · χA , implying
Z (
∞ for µ(A) > 0,
∞ · f dµ = lim (n · µ(A)) =
X n→∞ 0 for µ(A) = 0.
R R R
If µ(A) > 0, then,
R X
f dµ > 0 by (a), showing ∞ · X f dµ
R = ∞ = X∞ R · f dµ . If
µ(A) = 0, then X f dµ = 0 by (a). Thus, using (A.2e), ∞ · X f dµ = 0 = X ∞ · f dµ .
(c): If f ≤ g, then g − f ∈ M+ (A). Thus, (b) applies to g = f + (g − f ), yielding
Z Z Z Z
g dµ = f dµ + (g − f ) dµ ≥ f dµ ,
X X X X
as desired.
(d): While “≤” is immediate from Def. 2.4, “≥” is due to (c).
Theorem 2.7 (Monotone Convergence). If (X, A, µ) is a measure space and (fi )i∈N is
an increasing sequence in M+ (A), then
Z Z
lim fi dµ = lim fi dµ . (2.9)
X i→∞ i→∞ X
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 75
Proof. Letting f := limi→∞ fi , we have fi ↑ f and f ∈ M+ (A) by Th. 1.82(a). Now “≥”
in (2.9) is due to Lem. 2.6(c). To prove the remaining inequality, consider φ ∈ S + (A),
φ ≤ f , and β ∈]1, ∞[. Then, for each n ∈ N, we obtain Bn := {βfn ≥ φ} ∈ A and
φn := φ · χBn ∈ S + (A). Also, clearly, Bn ↑ X and φn ↑ φ. Then Lem. 2.5(a) yields
Z Z Z
φ dµ ≤ lim φ · χBn dµ ≤ β lim fn dµ
X n→∞ X n→∞ X
R R
and X φ dµ ≤ limn→∞ X fn dµ , since β > 1 was arbitrary. Now Lem. 2.6(d) yields
“≤” in (2.9) and completes the proof.
Corollary 2.8. If (X, A, µ) is a measure space and (gi )i∈N is a sequence in M+ (A),
then Z !
X∞ X∞ Z
gi dµ = gi dµ . (2.10)
X i=1 i=1 X
Pn
Proof. Letting, for each n ∈ N, fn := i=1 gi , (2.10) is precisely (2.9).
Example 2.9. (a) Theorem 2.7 does, in general, not hold without the monotonicity
assumption of the fn : Consider (X, A, µ) := (R, B 1 , β 1 ) with fn := n1 χ[0,n] . Then
fn → 0 (even uniformly), but
Z Z
1
fn dµ = · n = 1 6→ 0 = 0 dµ .
X n X
We are already familiar with the notation K to denote either R or C. Similarly, it will
sometimes be useful to treat R and C in parallel, giving rise to the following notation:
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 76
Notation 2.10. We use the symbol K̂ to denote either R or C; we write (K̂, B̂) to
denote either (R, B) or (C, B 2 ).
Definition 2.11. Let (X, A, µ) be a measure space and let f : X −→ K̂ be measurable.
(a) f is called integrable (or, more precisely, µ-integrable) if, and only if,
Z Z
+
(Re f ) dµ < ∞, (Re f )− dµ < ∞,
X X
Z Z (2.11)
(Im f )+ dµ < ∞, (Im f )− dµ < ∞.
X X
(a) A measurable function f : X −→ R is integrable if, and only if, both f + and f −
are integrable and, in that case,
Z Z Z
f dµ = +
f dµ − f − dµ .
X X X
(b) A measurable function f : X −→ C is integrable if, and only if, both Re f and
Im f are integrable and, in that case,
Z Z Z
f dµ = Re f dµ + i Im f dµ ,
X X X
Z Z Z Z
Re f dµ = Re f dµ , Im f dµ = Im f dµ .
X X X X
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 77
Lemma 2.13. Let (X, A, µ) be a measure space, let A ∈ A, and let f : X −→ K̂.
(b) It is f · χA ∈ M+ (A) if, and only if, f ↾A ∈ M+ (A|A), and, in that case,
Z Z Z
f dµ = f · χA dµ = f ↾A d(µ↾A|A ) (2.14)
A X A
(c) It is f · χA µ-integrable if, and only if, f ↾A is (µ↾A|A )-integrable, and, in that case,
(2.14) holds.
Since Ac ∈ A, we obtain (f · χA )−1 (B) ∈ A if, and only if, (f ↾A )−1 (B) ∈ A|A, proving
(a).
(b): Since f · χA ≡ 0 on Ac , the equivalence claimed in (b) is clear. If f · χA ∈ M+ (A),
then let (φn ) be a sequence in S + (A) with φn ↑ (f · χA ). Then (φn ↾A ) ↑ (f ↾A ) (with
φn ↾A ∈ S + (A|A)) and
Z Z Z Z
φn↾Ac ≡0
f dµ = f · χA dµ = lim φn dµ = lim (φn ↾A ) d(µ↾A|A )
A X n→∞ X n→∞ A
Z
= f ↾A d(µ↾A|A ),
A
proving (b).
(c) is a direct consequence of (b).
Theorem 2.14. Let (X, A, µ) be a measure space, f : X −→ K̂. Then the following
statements are equivalent:
(i) f is integrable.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 78
(v) f is measurable and there exists an integrable g ∈ M+ (A) such that |f | ≤ g (i.e.
|f | is dominated by g).
Proof. The equivalences of (i) – (iii) are immediate from Def. 2.11(a).
(iii) implies (iv) by setting p := (Re f )+ , q := (Re f )− , r := (Im f )+ , s := (Im f )− .
(iv) ⇒ (v): If f = p − q + i (r − s) with integrable p, q, r, s ∈ M+ (A), then |f | ≤ g :=
p + q + r + s, where g is integrable by Lem. 2.6(b).
R R
(v) implies (vi), since X |f | dµ ≤ X g dµ < ∞.
(vi) implies (iii): Let h ∈ {(Re f )+ , (Re f )− , (ImRf )+ , (Im f )R− }. Then h is measurable,
since f is measurable, and h is integrable, since X h dµ ≤ X |f | dµ < ∞.
Example 2.15. It can happen that |f | is integrable, but f is not: Let (X, A, µ) :=
([0, 1], B 1 , β 1 ), A ⊆ X, A ∈
/ B 1 . Then f := χA − χAc is not measurable (in particular,
not integrable), but |f | ≡ 1 is integrable.
Theorem 2.16. Let (X, A, µ) be a measure space, let f, g : X −→ K̂ be integrable, let
A, B ∈ A, and let α, β ∈ K.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 79
(f ) Mean Value Theorem for Integration: Let f be R-valued and assume there exist
numbers m, M ∈ R such that m ≤ f ≤ M on A. Moreover, let p : X −→ [0, ∞] be
such that χA p and χA f p are integrable. Then
Z Z Z
m p dµ ≤ f p dµ ≤ M p dµ . (2.15e)
A A A
where the theorem’s name comes from the fact that, for 0 < µ(A) < ∞,
Z Z
−1
− f dµ := µ(A) f dµ (2.15h)
A A
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 80
R f, g are C-valued,
If R then, since
R Re(f + g) = Re f + Re g and Im(f + g) = Im f + Im g,
X
(f + g) dµ = X
f dµ + X
g dµ R case. If f isR R-valued and
follows from the R-valued
+ + − −
α ≥ 0, then (αf ) = αf , (αf ) = αf and we obtain X αf dµ = α X f dµ due to
Lem. 2.6(b). If α < 0, then (αf )+ = |α|f − , (αf )− = |α|f + , and one computes
Z Z Z Z
− +
αf dµ = |α| f dµ − f dµ = α f dµ .
X X X X
Finally, using the R-valued case, one computes for C-valued f and α ∈ C,
Z Z Z
(αf ) = (Re α Re f − Im α Im f ) dµ + i (Re α Im f + Im α Re f ) dµ
X X X
Z Z Z Z
= Re α Re f dµ − Im α Im f dµ + i Re α Im f dµ + i Im α Re f dµ
X X X X
Z
=α f dµ ,
X
proving (b).
(c): Since f χA∪B = f χA + f χB , the case f ∈ M+ (A) follows from Lem. 2.6(b) and the
integrable case follows from (b).
(d): If f ≤ g, then g − f ≥ 0. Thus, (b) applies to g = f + (g − f ), yielding
Z Z Z Z
g dµ = f dµ + (g − f ) dµ ≥ f dµ ,
A A A A
as desired.
(e) follows from (d): Due to the existence of polar coordinates for complex numbers,
there exists ζ ∈ K with |ζ| = 1 and
Z Z Z
f dµ = ζ f dµ = ζf dµ .
A A A
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 81
proving (e).
(f): Since mp ≤ f p ≤ M p, we compute
Z (d)
Z (d)
Z
m p dµ ≤ f p dµ ≤ M p dµ .
A A A
R R
If A p dµ = 0, then
R (2.15e) implies A
f p dµ = 0 and (2.15f) is immediate. It remains
to consider that
R A
p dµ > 0. If f (A) = [m, M ], then (2.15e) shows there is ξ ∈ A such
f p dµ
that f (ξ) = RA , i.e. (2.15f) holds.
A p dµ
(g) follows from Th. 2.14(v), since max(f, g) and min(f, g) are measurable and both
| max(f, g)| and | min(f, g)| are dominated by |f | + |g|.
Here are some examples of potentially useful properties that might hold µ-almost ev-
erywhere: Functions f, g : X −→ Y are equal µ-a.e. if, and only if, there exists a µ-null
set N ∈ A such that f ↾N c = g ↾N c . A function f : X −→ R is finite µ-a.e. if, and only
if, there exists a µ-null set N such that f (N c ) ⊆ R. The sequence (fn )n∈N of functions
fn : X −→ K̂ converges a.e. to f : X −→ K̂ if, and only if, there exists a µ-null set N
such that limn→∞ fn ↾N c = f ↾N c .
Note that “P holds µ-a.e.” does not imply that M := {x ∈ X : P (x) is false} ∈ A. It
merely implies that there exists a µ-null set N ∈ A such that M ⊆ N (of course, if µ is
complete, then we do know M ∈ A).
Theorem 2.18. Let (X, A, µ) be a measure space.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 82
Proposition 2.19 (Fatou’s Lemma). Let (X, A, µ) be a measure space. Then, for each
sequence (fn )n∈N in M+ (A), one has
Z Z
lim fn dµ ≤ lim inf fn dµ . (2.19)
X n→∞ n→∞ X
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 83
we obtain Z Z Z
f dµ ≤ lim inf fk dµ = lim inf fn dµ ,
X n→∞ k≥n X n→∞ X
which establishes the case.
Proof. The fn are integrable by Th. 2.18(c) and then |f | = limn→∞ |fn | ≤ g µ-a.e.
implies f to be integrable as well. Then we know that outside of a µ-null set N ∈ A,
all the functions g, f, fn , n ∈ N, must be K-valued and limn→∞ fn = f . Since all the
involved integrals over N equal 0, we may assume, without loss of generality, that N = ∅.
Then, for each n ∈ N, gn := |f | + g − |fn − f | ∈ M+ (A), and Fatou (Prop. 2.19) yields
Z Z Z
(|f | + g) dµ = lim gn dµ ≤ lim inf gn dµ
X X n→∞ n→∞ X
Z Z
= (|f | + g) dµ − lim sup |fn − f | dµ .
X n→∞ X
R
Since X
(|f | + g) dµ ∈ R+
0 , we obtain (2.20b). As
Z Z Z
∀ f dµ − f dµ ≤ |fn − f | dµ ,
n∈N n
X X X
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 84
Note that we have already seen in Ex. 2.9(a), in the context of the monotone convergence
theorem, that one can not expect (2.20a) to hold without additional hypotheses (such
as monotonicity or domination).
Corollary 2.21. Let (X, A, µ) be a measure space.
P
(a) Let f, fn : X −→ K̂ be measurable functions, n ∈ N, such that f = ∞ n=1 fn µ-a.e.
+
Moreover,
Pn assume there exists an integrable g ∈ M (A) such that, for each n ∈ N,
| k=1 fk | ≤ g holds µ-a.e. Then all fn and f are integrable with
Z ∞ Z
X
f dµ = fn dµ .
X n=1 X
S∞
(b) Let f : X −→ K̂ be integrable and A, An ∈ A such that A = n=1 An and, for
k 6= l, µ(Ak ∩ Al ) = 0. Then
Z ∞ Z
X
f dµ = f dµ .
A n=1 An
P Pn−1
Proof. (a): The integrability of the fn follows since fn = nk=1 fk − k=1 fk , where both
sums
R on the right-hand side are integrable by hypothesis. Now the Pnclaimed formula for
X
f dµ is obtained from applying Th. 2.20 with fn replaced by k=1 fk .
S
(b): Since {|f | = ∞} and (k,l)∈N2 , k6=l (Ak ∩ Al ) are both µ-null sets, we may assume,
without loss of generality,Pthat A is the disjoint
Pn union of the PA n and that f · χA is
∞ n
K-valued. Then f · χA = n=1 f · χAn with | k=1 f · χAk | ≤ k=1 |f | · χAk ≤ |f | · χA ,
i.e. (b) now follows from (a).
Theorem 2.22. Let (Z, d) be a metric space and let (X, A, µ) be a measure space. Let
f : Z × X −→ K̂. Fix z0 ∈ Z and assume f has the following properties:
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 85
Proof. Let (zm )m∈N be a sequence in Bǫ (z0 ) such that limm→∞ zm = z0 . According to
(a), this gives rise to a sequence (gm )m∈N of integrable functions
Since, by (c), all gm are dominated by the integrable function h, we can apply the
dominated convergence Th. 2.20 to obtain
Z Z Z
Th. 2.20
lim φ(zm ) = lim f (zm , x) dµ(x) = lim gm (x) dµ(x) = g(x) dµ(x)
m→∞ m→∞ X m→∞ X X
Z
= f (z0 , x) dµ(x) = φ(z0 ),
X
proving φ is continuous at z0 .
Theorem 2.23. Let I ⊆ R be a nontrivial interval and let (X, A, µ) be a measure space.
Let f : I × X −→ K. Fix t0 ∈ I and assume f has the following properties:
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 86
Proof. Let (tm )m∈N be a sequence in I \ {t0 } such that limm→∞ tm = t0 . According to
(a), this gives rise to a sequence (gm )m∈N of integrable functions
f (tm , x) − f (t0 , x)
gm : X −→ K, gm (x) := ,
tm − t0
such that, using (b), limm→∞ gm = g µ-a.e., where
(
f ′ (t0 , x) if x 6∈ N ,
g : X −→ K, g(x) :=
0 if x ∈ N .
Since, by (c), all gm are dominated by the integrable function h, we can apply the
dominated convergence Th. 2.20 to obtain the integrability of the gm and g, as well as
Z Z
φ(tm , x) − φ(t0 , x) Th. 2.20
lim = lim gm (x) dµ(x) = g(x) dµ(x)
m→∞ tm − t0 m→∞ X X
Z
= f ′ (t0 , x) dµ(x) ,
X
Corollary 2.24. The statement of Th. 2.23 remains true if (b) and (c) are replaced by
(note that, this time, the null set, where the condition does not hold, must not
depend on t).
Proof. Since (b′ ) immediately implies (b), it remains to verify that (c′ ) implies (c). To
this end, let N be a µ-null set such that, for each x ∈ X \ N , |fx′ (t)| ≤ h(x) holds for
each t ∈ I. Then, for each x ∈ X \ N , by the mean value theorem of differential calculus
[Phi16a, Th. 9.18], for each t ∈ I \ {t0 }, there exists τ (t, x) ∈ I such that
f (t, x) − f (t0 , x) ′
= fx τ (t, x) ≤ h(x),
t − t0
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 87
We will now see in Th. 2.26(a) that every Riemann integrable function is also Lebesgue
integrable with identical values for the integral. Thus, the Lebesgue integral can be
seen as a generalization of the Riemann integral. In [Phi16a, Sec. 10], we only studied
the Riemann integral for functions on one-dimensional intervals. However, an analogous
construction can be performed for functions on intervals in Rn , n ∈ N. It is provided in
Sec. D of the Appendix. While the following Th. 2.26 is provided for arbitrary n ∈ N,
one can understand its contents and structure already from the case n = 1 (in particular,
without having studied Sec. D). The main difference between n = 1 and n > 1 is that
the notation gets somewhat more complicated.
As a consequence of Th. 2.26(a), for functions on R, one will, in most cases, still compute
Lebesgue integrals as Riemann integrals, using the techniques introduced in [Phi16a,
Sec. 10]. The most important technique for computing integrals on (subsets of) Rn is
the Fubini Th. 2.36(b) below, which, if it applies, allows to compute an integral of a
function on Rn by evaluating n one-dimensional integrals.
(b) f is Riemann integrable if, and only if, the set of points where f is discontinuous
is a λn -null set (i.e. if, and only if, f is continuous λn -a.e.).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 88
j · (bk − ak )2−N for each j ∈ {0, . . . , 2N }, into the disjoint intervals Ik,1 N
:= [xN N
k,0 , xk,1 ],
N N N N N N N
Ik,2 :=]xk,1 , xk,2 ], . . . , Ik,2N :=]xk,2N −1 , xk,2N ], and the resulting partition ∆ of I (cf.
Def. D.2). Analogous to Def. D.2, for each (k1 , . . . , kn ) ∈ P (∆N ) = {1, . . . , 2N }n , define
n
Y
I(jN1 ,...,jn ) := N
Ik,j k
,
k=1
which is as in (D.4), except that we are using the (disjoint) halfopen intervals here.
Then, analogous to Rem. D.3, we obtain the (now disjoint) unions
[˙
I= IpN .
p∈P (∆N )
we obtain Z Z
N n N
r(∆ , f ) = gN dλ , R(∆ , f ) = hN dλn .
I I
Moreover, clearly, gN ≤ f ≤ hN , (gN )N ∈N is increasing, and (hN )N ∈N is decreas-
ing. Thus, the pointwise limits g := limN →∞ gN and h := limN →∞ hN exist, are β n -
measurable, bounded, and, hence, Lebesgue integrable. They also satisfy g ≤ f ≤ h.
(a): Suppose f is Riemann integrable. Since
∀ |gN |, |hN | ≤ |g1 | + |h1 |,
N ∈N
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 89
R
Thus, I (h − g) dλn = 0, such that h = g λn -a.e. on I by Lem. 2.6(a). Since g ≤ f ≤ h,
this also yields f = g λn -a.e. on I. In consequence, there exists a λn -null set M ∈ Ln |I
such that f = g on M c . If B ∈ B 1 , then
proving (a).
(b): Let D := {x ∈ I : f not continuous at x} and let
[ [
R := ∂IpN
N ∈N p∈P (∆N )
be the set of boundary points occurring in any of the considered partitions of I, which,
as a countable union of λn -null sets is itself a λn -null set. We show that D ⊆ R∪{g < h}
by proving that (R∪{g < h})c = Rc ∩{g = h} ⊆ Dc : Let x ∈ Rc ∩{g = h} and let ǫ > 0.
Note f (x) = g(x) = h(x). Choose N ∈ N such that f (x)−ǫ < gN (x) < hN (x) < f (x)+ǫ.
Since x ∈/ R, there exists pN ∈ P (∆N ) such that x ∈ int(IpNN ). Set δ := dist(x, ∂IpN ).
Then δ > 0 and, for each y ∈ Rn with ky − xk < δ, one has y ∈ IpNN , implying
showing |f (y) − f (x)| < ǫ and the continuity of f at x, i.e. x ∈ Dc as claimed. Now,
if f is Riemann integrable, then, as shown in the proof of (a), g = h λn -a.e., i.e.
λn ({g < h}) = 0. Since D ⊆ R ∪ {g < h}, this implies λn (D) = 0. For the converse, we
show {g < h} ⊆ D: Indeed, for x ∈ I with g(x) < h(x) let δ := (h(x) − g(x))/3. Then,
for each ǫ > 0, there exist x1 , x2 ∈ Bǫ (x) such that f (x1 ) ≤ g(x)+δ and f (x2 ) ≥ h(x)−δ,
showing x ∈ D and {g < h} ⊆ D. Thus, if D ∈ Ln with λn (D) = 0, then g = h λn -a.e.,
yielding
Z Z
N
lim r(∆ , f ) = lim gN dλ = g dλn
n
N →∞ N →∞ I
Z Z I
= h dλn = lim hN dλn = lim R(∆N , f ),
I N →∞ I N →∞
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 90
is everywhere discontinuous and, thus, not Riemann integrable (cf. Ex. D.7(b) and
[Phi16a, Ex. 10.7(b)]). However, f is the characteristic
R function of the β n -null set
n n
I ∩ Q , i.e. f is Lebesgue integrable with I f dλ = 0. Now let (r1 , r2 , . . . ) be an
enumeration of I ∩ Qn and define
(
1 for x ∈ {r1 , . . . , rk },
∀ φk : I −→ R, φk (x) :=
k∈N 0 otherwise.
(b) The following example shows that there exist Riemann integrable functions that
are not Borel measurable: Let C ⊆ I := [0, 1] be the Cantor set of Def. and Rem.
1.61 and let A ⊆ C be such that A ∈/ B 1 (such an A exists since #P(C) = #P(R),
but #B 1 = #R). Then f : I −→ R, f := χA isR continuous on the open set I \ C.
Since λ1 (C) = 0, f is Riemann integrable with I f = 0.
—
While we saw in Th. 2.26(a) that every Riemann integrable function is Lebesgue in-
tegrable, for improperly Riemann integrable functions this only remains true if the
improper Riemann integral converges absolutely:
Theorem 2.28. Let a, b, c ∈ R, a < c < b. Let I ⊆]a, b[ be one of the following three
kinds of intervals: I = [c, b[, I =]a, c], or I =]a, b[. Moreover, assume f : I −→ R to be
locally Riemann integrable, i.e. Riemann integrable over each compact subinterval of I.
Then f is Lebesgue integrable if, and only if, |f | is improperly
R Riemann integrable. In
that case, both integrals agree, i.e. (2.26) holds, where R- now denotes the improper
Riemann integral.
Proof. Let m := inf I, M := sup I. Moreover, let (an )n∈N and (bn )n∈N be sequences in
I such that m < an < bn < M and an ↓ m as well as bn ↑ M . If, for each n ∈ N,
fn := f · χ[an ,bn ] , then fn ↑ f on ]m, M [. Since f is locally Riemann integrable, each
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 91
If |f | is improperly Riemann integrable, then the left-hand side is finite, i.e. |f | (and,
thus, f ) is Lebesgue integrable. Conversely, if f is Lebesgue integrable, then so is
|f |, i.e. the right-hand side is finite, implying |f | to be improperly Riemann integrable.
Finally, if |f | is improperly Riemann integrable, then so is f by [Phi16a, Prop. 10.39(b)],
implying, together with Th. 2.26(a) and the dominated convergence Th. 2.20,
Z M Z bn Z Z
R- f = lim R- f = lim f dλ = f dλ1 ,
1
m n→∞ an n→∞ [an ,bn ] I
(a) The gamma function is well-defined by (2.28), i.e. the integral exists as an improper
Riemann integral and as a Lebesgue integral with values in R+ : Since, for each
t, x > 0, one has e−t tx−1 > 0, and fx : R+ −→ R+ , fx (t) := e−t tx−1 , is continuous,
it suffices to show fx is dominated by some improperly Riemann integrable function
gx . In anticipation of (c) below, it will actually be useful to estimate fx uniformly
for each x in some bounded interval. Thus, let α, β ∈ R+ such that 0 < α < β.
Then, for each x ∈ [α, β],
∀ 0 < fx (t) ≤ tx−1 ≤ tα−1 ,
0<t≤1
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 92
dominates
R 1 α−1 fx for each x ∈ [α, β]. It is also improperly Riemann integrable, since
0
t dt = α1 by [Phi16a, Ex. 10.35(a)], and, if (tk )k∈N is a sequence in ]1, ∞[ such
that limk→∞ tk = ∞, then
Z ∞ Z tk
−t/2
e dt = lim e−t/2 dt
1 k→∞ 1
h itk
= lim − 2e−t/2 = lim (2e−1/2 − 2e−tk /2 ) = 2e−1/2 .
k→∞ 1 k→∞
Using [Phi16a, Ex. 9.5(b)], we differentiate the integrand of (2.28) k times with
respect to x to obtain, for each x ∈ R+ ,
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 93
Thus, according to Cor. 2.24, (2.31) is proved if we can estimate each integrand of
(2.31) by an integrable function, uniformly for x ∈ [α, β], where 0 < α < β as in
(a). For each k ∈ N and x ∈ [α, β], we estimate
∀ 0 < | ln t|k e−t tx−1 ≤ | ln t|k tα/2 t−α/2 tx−1 ≤ Cα,k tα/2−1 ,
0<t≤1
∀ 0 < | ln t|k e−t tx−1 = | ln t|k tx−1 e−t/2 e−t/2 ≤ Mβ,k e−t/2 ,
t≥1
dominates | ln t|k e−t tx−1 for each x ∈ [α, β]. That it is improperly Riemann inte-
grable follows analogously to the integrability of gα,β in (a).
and
Z ∞
1 √ −x2
e dx = Γ = π: (2.33)
−∞ 2
P∞ xi x
For each x ∈ R, one has 1 + x ≤ i=0 i! = e . Thus, for each x ∈] − 1, ∞[,
applying ln yields ln(1 + x) ≤ x. If 0 ≤ t < n ∈ N, then −1 < − nt and we can
apply ln(1 + x) ≤ x with x := − nt to obtain (1 − nt )n ≤ e−t . In consequence, we
can apply the dominated convergence Th. 2.20 to obtain, for each x ∈ R+ ,
Z ∞ Z ∞ n
−t x−1 t
Γ(x) = e t dt = lim 1− tx−1 · χ]0,n[ dt
0 n→∞ 0 n
Z n n
t
= lim 1− tx−1 dt ,
n→∞ 0 n
proving the first identity of (2.32). To obtain the second identity of (2.32), for each
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 94
= ...
Z n 0
n! t
=0+ n 1− tx+n−1 dt
n x(x + 1) · · · (x + n − 1) 0 n
n! h in nx n!
= n tx+n = .
n x(x + 1) · · · (x + n) 0 x(x + 1) · · · (x + n)
The first identity in (2.33) can be proved using the change of variables t := x2 :
Z ∞ Z ∞ −t
−x2 e −1/2 1 1
e dx = t dt = Γ .
0 0 2 2 2
The second identity in (2.33) is a consequence of (2.32) and the so-called Wallis
product (see Prop. E.1(c) of the Appendix for an elementary proof of the Wallis
product)
∞ n
π Y 2k 2k Y 2k 2k 2 2 4 4 6 6
= · = lim · = · · · · · ··· :
2 k=1 2k − 1 2k + 1 n→∞
k=1
2k − 1 2k + 1 1 3 3 5 5 7
(2.34)
v
u n
√ u Y 2k 2k
π = lim t2 (2n + 1) ·
n→∞
k=1
2k + 1 2k + 1
r √ !
2 (2n + 1) n n! 2n
= lim ·
n→∞ n 1 · 3 · · · (2n + 1)
√
n n! 2n+1 1
= lim =Γ ,
n→∞ 1 · 3 · · · (2n + 1) 2
where we also used the continuity of the square root function.
2.5 Products
2.5.1 Product Measure
n
Qn 2.30. Let nn ∈ N and let (Xi , Ai , µi )i=1 be a finite family of measure spaces,
Definition
X := i=1 Xi , A := ⊗i=1 Ai . Then a measure µ on (X, A) is called a product measure
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 95
Lemma 2.31. QnLet n ∈ N, n ≥ 2, and let (Xi , Ai )ni=1 be a finite family of measurable
n
spaces, X := i=1 Xi , A := ⊗i=1 Ai . Define the sections
( n−1
)
Y
∀ ∀ Mxn := (x1 , . . . , xn−1 ) ∈ Xi : (x1 , . . . , xn ) ∈ M . (2.36)
M ⊆X xn ∈Xn
i=1
(a) Forming sections commutes with set-theoretic rules: Let M ⊆ X and let (Mj )j∈J
be a family of subsets of X, J 6= ∅. Also let Ei ⊆ Xi for i ∈ {1, . . . , n}. Then, for
each xn ∈ Xn ,
!
\ \
Mj = (Mj )xn , (2.37a)
j∈J j∈J
!x n
[ [
Mj = (Mj )xn , (2.37b)
j∈J xn j∈J
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 96
proving (2.37e).
(b): Let
M := {M ⊆ X : Mxn ∈ Q for each xn ∈ Xn }.
We show M to e a σ-algebra: It is ∅ ∈ Q, since ∅ ∈ An . Let M ∈ M and xn ∈ Xn .
Then, by (2.37c), (M c )xn = (Mxn )c ∈ Q, since MxSn ∈ Q. Thus, S M c ∈ M. Let (Mk )k∈N
be a sequence in M, xn ∈SXn . Then, by (2.37b), ( k∈N Mk )xn = k∈N (Mk )xn ∈ Q, Q since
each (Mk )xn ∈ Q. Thus, k∈N Mk ∈ M, showing M to be a σ-algebra. If M = ni=1 Ai
with each Ai ∈ Ai , then (2.37d) implies M ∈ M. Thus, A ⊆ M, since M is a σ-algebra.
Since A ⊆ M is precisely the claim of (b), the proof is done.
(c): If B ∈ B, then, by (2.37e), f (·, xn )−1 (B) = f −1 (B) xn ∈ Q by (b), since f −1 (B) ∈
A by the measurability of f .
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 97
n
Qn 2.32. Let nn∈ N and let (Xi , Ai , µi )i=1 be a finite family of measure spaces,
Theorem
X := i=1 Xi , A := ⊗i=1 Ai .
(a) There always exists at least one product measure on (X, A).
(b) Let n = 2. If µ1 is σ-finite, then, using the notation of (2.36), one has
and Z
µ : A −→ [0, ∞], µ(M ) := µ1 (Mx2 ) dµ2 (x2 ), (2.38b)
X2
defines a product measure.
(c) Let n ≥ 2. If each measure µi is σ-finite, then there exists a unique product measure
µ on (X, A), denoted by ⊗ni=1 µi := µ. Moreover, ⊗ni=1 µi is itself σ-finite and, using
the notation of (2.36), one has
n−1
∀ xn 7→ ⊗i=1 µi (Mxn ) is An -B-measurable (2.39a)
M ∈A
and Z
n−1
∀ (⊗ni=1 µi ) (M ) = ⊗i=1 µi (Mxn ) dµn (xn ). (2.39b)
M ∈A Xn
Qn−1 Qn
In terms of the canonical identification ( i=1 Xi ) × Xn ∼
= i=1 Xi , one has
n−1
(⊗i=1 µi ) ⊗ µn = ⊗ni=1 µi . (2.39c)
Proof. (a): For n = 1, there is nothing to prove. We consider the case n = 2, from
which the case n > 2 follows by induction. Let n = 2. We know from Prop. 1.17(a) that
E := A1 ∗ A2 = {A1 × A2 : Ai ∈ Ai , i = 1, 2} is a semiring. We also know A = σ(E).
We now define µ on E by (2.35), i.e. µ(A1 × A2 ) := µ1 (A1 )µ2 (A2 ) for A1 ∈ A1 , A2 ∈ A2 .
We show that this defines a premeasure on E: µ(∅) = 0 is clear. Let A ∈ A1 , B ∈ A2 ,
and let (Ai )i∈N and (Bi )i∈N be sequences in A1 and A2 , respectively, such that A × B is
Ṡ
the disjoint union of the Ai × Bi , i.e. A × B = i∈N (Ai × Bi ). Then
Z Z
(2.37d)
µ(A × B) = µ1 (A) χB (x) dµ2 (x) = µ1 (A × B)x dµ2 (x)
X2 X2
Z ! Z X ∞
(2.37b) [
= µ1 (Ai × Bi )x dµ2 (x) = µ1 (Ai × Bi )x dµ2 (x)
X2 i∈N X2 i=1
∞ Z
X ∞
X
Cor. 2.8
= µ1 (Ai × Bi )x dµ2 (x) = µ(Ai × Bi ),
i=1 X2 i=1
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 98
M := {M ∈ A : fM is measurable},
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 99
This implies
1
∀ β (Ak ) = 0 ∨ Bk finite .
k∈N
S S
Thus, letting A := {Ak : β 1 (Ak ) = 0}, B := {Bk : Bk finite}, we have D ⊆ (A×R)∪
(R × B). Note π1 (M ) = π2 (M ) for each M ⊆ D. We obtain π1 (D \ (A × R)) = [0, 1] \ A
and π1 (D \ (A × R)) = π2 (D \ (A × R)) ⊆ B. Since β 1 ([0, 1] \ A) = 1, but β 1 (B) = 0,
this is a contradiction, proving α(D) = ∞. Now let β be the product measure given by
Th. 2.32(b). Then
Z Z Z
1 1
β(D) = β (Dx ) dµ2 (x) = β ({x}) dµ2 (x) = 0 dµ2 (x) = 0.
R [0,1] [0,1]
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 100
It is shown in [Beh87, pp. 94–96] that, even though µ2 is not σ-finite, in our considered
situation, for each M ∈ B 2 , the map x 7→ µ2 (Mx ), Mx := {y ∈ R : (x, y) ∈ M } is
λ1 -measurable and then
Z
2
γ : B −→ [0, ∞], γ(M ) := µ2 (Mx ) dλ1 (x),
R
(c) Let n ∈ N. We compute the volume of n-dimensional balls with respect to the
Euclidean norm: Let x ∈ Rn , r > 0, Br (x) := {y ∈ Rn : ky − xk2 < r}.
π n/2
vol(Br (x)) := β n (Br (x)) = ωn rn , where ωn := β n (B1 (0)) = , (2.41)
Γ( n2 + 1)
where Γ denotes the Gamma function from Ex. 2.29: It suffices to prove the formula
for ωn : Indeed, consider the bijective affine map A : Rn −→ Rn , A(v) := rv + x.
Then A(B1 (0)) = Br (x), det A = rn , and β n (Br (x)) = rn β n (B1 (0)) by (1.47b).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 101
We now prove the formula for √ ωn via induction on n ∈ N: For n = 1, the claim
1 π 3 1 1 1√
is ω1 = β (] − 1, 1[) = 2 = Γ( 3 ) , which holds, as Γ( 2 ) = 2 Γ( 2 ) = 2 π by (2.29a)
2
and (2.33). For the induction step, let n > 1, write Rn = Rn−1 × R and use
β n = β n−1 ⊗ β 1 . For x ∈ R, we obtain the sections
( n−1
) (
X ∅ for |x| ≥ 1,
B1 (0) x = y ∈ Rn−1 : x2 + yi2 < 1 = n−1
i=1
B√1−x2 (0) for −1 < x < 1,
n−1
where B√ 1−x2
(0) denotes a ball in Rn−1 . Thus,
Z 1 Z 1
(2.39b) n−1
1
ωn = β B1 (0) x
dβ (x) = ωn−1 (1 − x2 )(n−1)/2 dx .
−1 −1
Rπ (E.3a)
In particular, for n = 2, ω2 = ω1 0 (sin t)2 dt = 2 · π2 = π = π
Γ(2)
, proving (2.41)
for n = 2. For n > 2,
Z π Z π
n−1
ωn = ωn−2 (sin t) dt (sin t)n dt .
0 0
Finally, using the induction hypothesis and (E.3), we obtain, for n even
n/2−1 n/2
π (n−2)/2 Y 2k Y 2k − 1 2π n/2 (2.29a) π n/2
ωn = 2π = =
Γ( n2 ) k=1
2k + 1 k=1 2k nΓ( n2 ) Γ( n2 + 1)
For the actual computation of higher-dimensional integrals, the theorems of Tonelli and
Fubini, which we provide in Th. 2.36 below, are among the most powerful tools, as they
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 102
and observes that this definition does not depend on the choice of z ∈ f (X) (since N
is a µ-null set). In practice, one will usually still denote g by f and one often does not
even specify z explicitly.
Theorem 2.36. Let nQ∈ N, n ≥ 2. For each i ∈ {1, . . . , n}, let (Xi , Ai , µi ) be a σ-finite
n
measure space, X := i=1 Xi . We consider two cases:
(ii) Each µi is complete and (X, A, µ) is the completion of the corresponding product
measure space.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 103
are A1 -measurable and A2 -measurable, respectively (in Case (ii), they are defined
according to Convention 2.35), and
Z
f (x1 , x2 ) dµ(x1 , x2 )
X1 ×X2
Z Z
= f (x1 , x2 ) dµ1 (x1 ) dµ2 (x2 )
X2 X1
Z Z
= f (x1 , x2 ) dµ2 (x2 ) dµ1 (x1 ). (2.42b)
X1 X2
where one can also arbitrarily permute the order of the inner integrals without
changing the overall value (as above, the inner integrals are defined according to
Convention 2.35 in Case (ii)).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 104
where π is a permutation of {1, . . . , n}, is finite, then all the integrals (for an
arbitrary permutation π of {1, . . . , n}) are finite, f is µ-integrable and, in particular,
all the assertions of (b) hold true.
Proof. It suffices to prove the case n = 2 – then the case n > 2 follows by induction
(and using that every permutation is a finite composition of transpositions).
(a): Case (i): Let M ∈ A = A1 ⊗ A2 . If x2 ∈ X2 , then, for the section Mx2 ⊆ X1 , one
has Mx2 ∈ A1 by Lem. 2.31(b), and x2 7→ µ1 (Mx2 ) is A2 -B-measurable by Th. 2.32(b).
One computes
Z Z
(2.38b)
χM (x1 , x2 ) dµ(x1 , x2 ) = µ(M ) = µ1 (Mx2 ) dµ2 (x2 )
X X2
Z Z Z Z
= χMx2 (x1 ) dµ1 (x1 ) dµ2 (x2 ) = χM (x1 , x2 ) dµ1 (x1 ) dµ2 (x2 )
X2 X1 X2 X1
N
X Z Z
= αk χMk (x1 , x2 ) dµ1 (x1 ) dµ2 (x2 )
k=1 X2 X1
Z Z
= f (x1 , x2 ) dµ1 (x1 ) dµ2 (x2 ) , (2.44)
X2 X1
proving the first equality of (2.42b) for nonnegative simple functions. Next, we consider
f ∈ M+ (A) and choose a sequence (φk )k∈N in S + (A) such that φk ↑ f . According to
Lem. 2.31(c), for each x2 ∈ X2 , f (·, x2 ) ∈ M+ (A1 ) and, for each k ∈ N, φk (·, x2 ) ∈
S + (A1 ). Since φk (·, x2 ) ↑ f (·, x2 ), we obtain, according to (2.4)
Z Z
φk (x1 , x2 ) dµ1 (x1 ) ↑ f (x1 , x2 ) dµ1 (x1 ) . (2.45)
X1 X1
R
Moreover, each function x2 7→ X1 φk (x1 , x2 ) dµ1 (x1 ) on the left-hand side of (2.45) is
R
A2 -measurable, implying the function x2 7→ X1 f (x1 , x2 ) dµ1 (x1 ) on the right-hand side
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 105
MCT
= lim φk (x1 , x2 ) dµ1 (x1 ) dµ2 (x2 )
X2 X1 k→∞
Z Z
(2.4)
= f (x1 , x2 ) dµ1 (x1 ) dµ2 (x2 ) , (2.46)
X2 X1
proving the first equality of (2.42b) for each f ∈ M+ (A) as claimed. To prove the
second equality of (2.42b) as well as all the remaining claims of (a) (in Case (i)), one
merely exchanges the roles of (X1 , A1 , µ1 ) and (X2 , A2 , µ2 ) in the above argument.
Case (ii): Let M ∈ A = (A1 ⊗ A2 )∼ . Then there exist A, C ∈ A1 ⊗ A2 and N ⊆ C, such
that (µ1 ⊗ µ2 )(C) = 0 and M = A ∪ N . If x2 ∈ X2 , then, for the section Mx2 ⊆ X1 , one
has Mx2 = Ax2 ∪ Nx2 , Nx2 ⊆ Cx2 . Since (µ1 ⊗ µ2 )(C) = 0, (2.38b) implies µ1 (Cx2 ) = 0
for µ2 -a.e. x2 ∈ X2 . Thus, since µ1 is complete, Mx2 ∈ A1 and χM (·, x2 ) = χMx2 is
µ1 -measurable for µ2 -a.e. x2 ∈ X2 . Using Convention 2.35, one computes
Z Z
(2.38b)
χM (x1 , x2 ) dµ(x1 , x2 ) = µ(M ) = µ(A) = µ1 (Ax2 ) dµ2 (x2 )
X X2
Z Z Z Z
= χAx2 (x1 ) dµ1 (x1 ) dµ2 (x2 ) = χM (x1 , x2 ) dµ1 (x1 ) dµ2 (x2 ) ,
X2 X1 X2 X1
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 106
implying Z
|f (x1 , x2 )| dµ1 (x1 ) < ∞ µ2 -a.e.
X1
In other words, f (·, x2 ) is µ1 -integrable for each x2 ∈ N c , where N is a µ2 -null set, i.e.,
letting f1 := (Re f )+ , f2 := −(Re f )− , f3 := i (Im f )+ , f4 := −i (Im f )− ,
Z 4 Z
X
∀ f (x1 , x2 ) dµ1 (x1 ) = fk (x1 , x2 ) dµ1 (x1 ) .
x2 ∈N c X1 X1
k=1
R
Moreover, for each k ∈ {1, 2, 3, 4}, x2 7→ X1 fk (x1 , x2 ) dµ1 (x1 ) (extended to N by 0) is
µ2 -integrable, since
Z Z Z Z
|fk (x1 , x2 )| dµ1 (x1 ) dµ2 (x2 ) ≤ |f (x1 , x2 )| dµ1 (x1 ) dµ2 (x2 ) < ∞.
X2 X1 X2 X1
X4 Z Z
= fk dµ = f dµ ,
k=1 X X
i.e. the first equality of (2.42b). Exchanging the roles of (X1 , A1 , µ1 ) and (X2 , A2 , µ2 )
completes the proof of (b).
(c) is just (a) combined with (b).
The technique employed in the proof of Th. 2.36 can often be employed to prove an asser-
tion about integrable functions: One first proves the assertion for characteristic functions
of measurable sets, then for simple functions (linear combinations of characteristic func-
tions of measurable sets), then for nonnegative measurable functions (pointwise limits
of simple functions), then, finally, for integrable functions f (whose parts (Re f )± and
(Im f )± are nonnegative measurable).
R (a) RLet I 3:= [1, 2] × [2, 3] × [0, 2]. We use the Fubini theorem to
Example 2.37.
compute I f := I f dλ for
2z
f : I −→ R, f (x, y, z) := :
(x + y)2
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 107
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 108
Observe that, on X,
x 1 x x
∂y arctan =− 2 · = − ,
y 1 + xy2 y 2 x2 + y 2
x x2 + y 2 − 2x2 x2 − y 2
∂x ∂y arctan = − = 2 ,
y (x2 + y 2 )2 (x + y 2 )2
y y
∂x arctan = − 2 ,
x x + y2
y y 2 − x2
∂y ∂x arctan = 2 .
x (x + y 2 )2
Thus,
Z 1Z 1 Z 1 y=1 Z 1
y 1 π
f (x, y) dy dx = dx = dx = [arctan x]10 = ,
0 0 0 x + y2
2
y=0 0 1+x
2 4
Z 1 Z 1 Z 1 x=1 Z 1
x 1 π
f (x, y) dx dy = − dy = − dy = − .
0 0 0 x2 + y2 x=0 0 1+y
2 4
2
In particular, f cannot be β -integrable over X.
2.6 Lp -Spaces
2.6.1 Definition, Norm, Completeness
and
N∞ (f ) := inf sup |f (x)| : x ∈ X \ N : N ∈ A, µ(N ) = 0 ∈ [0, ∞], (2.48b)
where the number N∞ (f ) is also known as the essential supremum of f (in (2.48b), it is
useful to set sup ∅ := inf[0, ∞] = 0 – this setting is only relevant if µ(X) = 0, resulting
in N∞ (f ) = 0). Note that |f | ≤ N∞ (f ) µ-a.e., since, for N∞ (f ) < ∞,
[ 1
|f | > N∞ (f ) = |f | > N∞ (f ) +
n∈N
n
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 109
is a µ-null set.
Theorem 2.39 (Hölder Inequality). Let (X, A, µ) be a measure space and p, q ∈ [1, ∞]
such that p1 + 1q = 1 (note 1/∞ = 0). If f, g : X −→ K̂ are measurable, then
N1 (f g) ≤ Np (f ) Nq (g), (2.49)
where (2.49) is known as Hölder inequality. We point out the following special cases:
(a) For p = q = 2, one obtains the Cauchy-Schwarz inequality. (b) If µ is counting
measure on {1, . . . , n}, n ∈ N, then one recovers the Hölder inequality of [Phi16b, Th.
1.12]. (c) If µ is counting measure on N, then one obtains the Hölder (and, in particular,
Cauchy-Schwarz) inequality for series of real (or complex) numbers.
Proof. If p = 1, q = ∞, then
Z Z
N1 (f g) = |f g| dµ ≤ N∞ (g) |f | dµ
X X
proves (2.49), and the case p = ∞, q = 1 is analogous. Thus, let 1 < p, q < ∞. If
Np (f ) = 0 or Nq (g) = 0, then f g = 0 µ-a.e., implying N1 (f g) = 0 and (2.49). As (2.49)
is also immediate for Np (f ) Nq (g) = ∞, it merely remains to consider the case 0 <
Np (f ), Nq (g) < ∞. Recall that, according to [Phi16a, Th. 9.35], if n ∈ N, x1 , . . . , xn ≥ 0
and λ1 , . . . , λn > 0 such that λ1 + · · · + λn = 1, then xλ11 · · · xλnn ≤ λ1 x1 + · · · + λn xn .
p q
Using this with n = 2, λ1 := p1 , λ2 := 1q , x1 := |f |/Np (f ) , x2 := |g|/Nq (g) , yields
p q
|f | |g| 1 |f | 1 |g|
≤ +
Np (f ) Nq (g) p Np (f ) q Nq (g)
(which, trivially, also holds for |f | = ∞ or |g| = ∞). Integrating this inequality over X
proves (2.49).
As in the special case of Kn in Analysis II, we obtain the Minkowski inequality (triangle
inequality for Np ) from the Hölder inequality:
Theorem 2.40. Let (X, A, µ) be a measure space and let f, g : X −→ K̂ be measurable
functions.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 110
If p = ∞, then
∀ sup |f (x) + g(x)| : x ∈ A ≤ sup |f (x)| : x ∈ A + sup |g(x)| : x ∈ A ,
A⊆X
proves N∞ (f + g) ≤ N∞ (f ) + N∞ (g). For the remaining case, consider 1 < p < ∞ and
p
f, g such that Np (f + g) > 0 as well as 0 ≤ Np (f ), Np (g) < ∞. Set q := (1 − p1 )−1 = p−1
and apply the Hölder inequality (2.49) to obtain
Z Z Z
p p p−1
Np (f + g) = |f + g| dµ ≤ |f | |f + g| dµ + |g| |f + g|p−1 dµ
X X X
(2.49)
≤ Np (f ) + Np (g) Nq |f + g|p−1
Z 1/q
(p−1)q
= Np (f ) + Np (g) |f + g| dµ
X
(p−1)q=p p/q
= Np (f ) + Np (g) Np (f + g)
p/q=p−1 p−1
= Np (f ) + Np (g) Np (f + g) . (2.52)
Next, we use p
|f + g|p ≤ 2 max(|f |, |g|) ≤ 2p |f |p + |g|p
and Np (f ), Np (g) < ∞ to conclude Np (f + g) ∈ R+ . Thus, we can divide (2.52) by
p−1
Np (f + g) to prove (2.50).
(b): For p = 1, (2.51) is just (2.50). Thus, let 0 < p < 1. Fix a ∈ R+ and consider the
function
φ : R+0 −→ R, φ(t) := ap + tp − (a + t)p ,
which is, clearly, continuous and, for t > 0, differentiable. Moreover,
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 111
(since p − 1 < 0, it is tp−1 > (a + t)p−1 ). Thus, φ is strictly increasing and, noting
φ(0) = 0, we have ap + tp − (a + t)p ≥ 0 for each t ≥ 0. We restate our findings as
∀ (a + b)p ≤ ap + bp .
a,b∈R+
0
Applying this with a := |f | and b := |g| yields (2.51a) after an integration over X. To
prove (2.51b), we, once again, fix a ∈ R+ , this time considering the function
ψ : R+
0 −→ R, ψ(t) := (a1/p + t1/p ) (a + t)−1/p ,
showing ψ to be strictly decreasing on [0, a] and strictly increasing on [a, ∞[ with a strict
1
global min at t = a, ψ(a) = 2 a1/p 2−1/p a−1/p = 21− p . Taking reciprocals in ψ(b) ≥ ψ(a)
proves
1
∀ (a + b)1/p ≤ 2 p −1 (a1/p + b1/p ).
a,b∈[0,∞]
(2.51a) 1/p 1
Np (f + g) ≤ Npp (f ) + Npp (g) ≤ 2 p −1 Np (f ) + Np (g) ,
proving (2.51b).
Definition and Remark 2.41. Let (X, A, µ) be a measure space and p ∈]0, ∞].
(a) Let Lp := LpK := LpK (µ) := LpK (X, A, µ) denote the set of all µ-measurable functions
f : X −→ K such that Np (f ) < ∞ (i.e. such that |f |p is µ-integrable (for p < ∞)).
For p ≥ 1, it is common to call Np the p-norm (caveat: on Lp , Np is, in general,
only a seminorm, see below) and to introduce the notation
∀ ∀ kf kp := Np (f ). (2.53a)
p∈[1,∞] f ∈Lp
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 112
We also define
∀ dp : Lp × Lp −→ R+
0, dp (f, g) := kf − gkp , (2.53b)
p∈[1,∞]
∀ dp : Lp × Lp −→ R+
0, dp (f, g) := Npp (f − g). (2.53c)
p∈]0,1[
Clearly, each Lp forms a vector space over K (here we use that Lp contains only
K-valued functions and not K̂-valued functions). Moreover, for p ∈ [1, ∞], Np
constitutes a seminorm on Lp (i.e. Np (0) = 0, Np is homogeneous of degree 1, and Np
satisfies the triangle inequality due to the Minkowski inequality (2.50), cf. [Phi16b,
Def. E.2]) and, in consequence, dp forms a pseudometric on Lp (i.e. dp (f, f ) = 0,
dp is symmetric, and dp satisfies the triangle inequality, cf. [Phi16b, Def. E.1]). For
p ∈]0, 1[, dp with the modified definition (2.53c) still forms a pseudometric on Lp :
Indeed, dp (f, f ) = 0 and dp (f, g) = dp (g, f ) are clear,
(b) Let N denote the set of measurable f : X −→ K that vanish µ-almost everywhere.
Clearly, N is a vector space over K, namely a subspace of Lp . Thus, we can define
the factor vector space (a.k.a. quotient vector space):
f ∼g :⇔ f −g ∈N ⇔ dp (f, g) = 0,
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 113
∀ k · k : Lp −→ R+
0, k[f ]kp := kf kp ,
p∈[1,∞]
∀ dp : Lp × Lp −→ R+
0, dp ([f ], [g]) := dp (f, g).
p∈]0,∞]
An obvious question is if, for p < q, one of the spaces Lp and Lq contains the other?
The answer is “no” in general (e.g., f ≡ 1 is in L∞ (λ1 ), but not in L1 (λ1 ), g with
g(t) := |t|−1/2 · χ]0,1] (t) is in L1 (λ1 ), but not in L∞ (λ1 )), but the answer is affirmative
for spaces with finite measure. The latter result is a simple consequence of the Hölder
inequality (2.49) and is provided as Th. 2.42. The complete answer is then given (with
proof omitted) in Rem. 2.43 below.
Theorem 2.42. Let (X, A, µ) be a finite measure space, i.e. µ(X) < ∞. Then, for each
p, q ∈]0, ∞] with 0 < p < q ≤ ∞, one has Lq ⊆ Lp and
1 1
∀q Np (f ) ≤ µ(X) p − q Nq (f ).
f ∈L
Proof. It suffices to prove the estimate, which, clearly, implies Lq ⊆ Lp . Thus, let
f ∈ Lq . For q = ∞, we compute
Z p1 Z p1
p
p 1
Np (f ) = |f | dµ ≤ N∞ (f ) dµ = N∞ (f ) µ(X) p ,
X X
thereby establishing the case. For q < ∞, set r := pq > 1 and s := (1 − 1r )−1 = r−1r
> 1.
1 1 p r pr q 1 s
Then r + s = 1. Since |f | ∈ L (since |f | = |f | ∈ L ) and 1 ∈ L , we obtain
Z (2.49)
Z r1
1 p
p
Np (f ) = p p
|f | · 1 dµ ≤ Nr (|f | ) Ns (1) = pr
|f | dµ µ(X) s = Nqp (f ) µ(X)1− q ,
X X
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 114
Remark 2.43. Let (X, A, µ) be a measure space. In generalization of Th. 2.42, one
can show (cf. Problems 6.2.5 and 6.2.6 in [Els07]) the equivalence of the statements
(iii) for each p, q ∈]0, ∞] such that 0 < p < q ≤ ∞, one has Lq ⊆ Lp ;
(iii′ ) for each p, q ∈]0, ∞] such that 0 < p < q ≤ ∞, one has Lp ⊆ Lq .
In particular, a space with µ(X) = ∞ can only have the properties (i) – (iii) if it
is not σ-finite. As a concrete simple example, consider X := {0, 1}, A := P(X),
µ(∅) := µ({0}) := 0, µ({1}) := µ(X) := ∞. Then every function f : X −→ K
is measurable, properties (ii) and (ii′ ) are clearly satisfied and, for each p ∈]0, ∞],
Lp = {(f : X −→ K) : f (1) = 0}.
—
The following Th. 2.44(a), regarding the completeness of the Lp spaces, is sometimes
known as the Riesz-Fischer theorem. However, part of the literature reserves the name
for a different theorem and, thus, we leave Th. 2.44(a) unnamed.
(a) All the spaces LpK (µ) are complete (with respect to the pseudometric dp ), where the
definition of Cauchy sequence and completeness in pseudometric spaces is precisely
the same as in metric spaces. In particular, all LpK (µ), p ∈ [1, ∞], are Banach
spaces; all LpK (µ), p ∈]0, 1[, are complete metric spaces.
(b) If (fk )k∈N is a sequence in LpK (µ), converging to g ∈ LpK (µ), then (fk )k∈N has a
subsequence that converges to g pointwise µ-almost everywhere. For p = ∞, (fk )k∈N
itself converges to g, even uniformly µ-almost everywhere.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 115
Proof. (a): Let (fk )k∈N be a Cauchy sequence in Lp . First, consider 0 < p < ∞. Then
there exists a subsequence (fkl )l∈N such that
TheP monotone convergence Th. 2.7 allows to take the limit for nP→ ∞, resulting in
Np ( ∞l=1 |gl |) ≤ 1. In particular, for µ-a.e. x ∈ X, the series
∞
l=1 gl (x) converges
absolutely and, in consequence, converges to some h(x) ∈ K. However,
n
X
∀ ∀ gl (x) = fk1 (x) − fkn+1 (x),
x∈X n∈N
l=1
showing there exists a µ-null set N such that (fkl )l∈N converges on N c (i.e. µ-a.e.) to
the measurable function
(
fk1 (x) − h(x) for x ∈ N c ,
f : X −→ K, f (x) :=
0 for x ∈ N .
We still need to verify f ∈ Lp and limk→∞ dp (fk , f ) = 0. To this end, let ǫ > 0. Then
there exists k0 ∈ N such that Npp (fk − fm ) < ǫ for each k, m > k0 . Fixing m > k0 and
applying Fatou (Prop. 2.19) to the sequence |fkl − fm |p l∈N yields
Z Z Z
p p p
Np (fm − f ) = |f − fm | dµ = lim |fkl − fm | dµ ≤ lim inf |fkl − fm |p dµ
X X l→∞ l→∞ X
If x ∈ N c , then (fk (x))k∈N is a Cauchy sequence in K (since (fk )k∈N is a Cauchy sequence
in L∞ ). Thus, there exists f (x) ∈ K such that f (x) = limk→∞ fk (x). On N c , the
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 116
convergence fk → f is even uniform: Given ǫ > 0, choose n0 ∈ N such that N∞ (fk −fl ) <
ǫ
2
for each k, l > n0 . Also choose m > n0 such that |fm (x) − f (x)| < 2ǫ . Then, for each
k > n0 (and independently of x ∈ N c ), |fk (x)−f (x)| ≤ |fk (x)−fm (x)|+|fm (x)−f (x)| <
ǫ
2
+ 2ǫ = ǫ. Extending f by 0 to X, we obtain f = limk→∞ fk χN c . In particular, f is
measurable as the pointwise limit of measurable functions. The uniform convergence
fk → f on N c also yields limk→∞ N∞ (fk − f ) = 0 and N∞ (f ) ≤ N∞ (f − fk ) + N∞ (fk ) <
∞, showing f ∈ L∞ .
(b): If (fk )k∈N converges in Lp , then it is a Cauchy sequence in Lp . Thus, the proof
of (a) shows the statement of (b) to be true if g is replaced by the function f ∈ Lp
constructed in the proof of (a). Then dp (f, g) ≤ dp (f, fk ) + dp (fk , g) → 0 for k → ∞,
showing dp (f, g) = 0, i.e. f = g µ-a.e., thereby establishing the case.
In Th. 2.44(b), for p < ∞, one can, in general, not expect for the sequence (fk )k∈∞ to
converge pointwise µ-a.e. to some f ∈ Lp , as is demonstrated by the following example:
Consider the measure space ([0, 1], B 1 , β 1 ) and the sequence of intervals
1 1 1 1 2 2 1
(Ik )k∈N := [0, 1], 0, , , 1 , 0, , , , , 1 , 0, , ... .
2 2 3 3 3 3 4
If, for each k ∈ N, fk := χIk , then, for each p ∈]0, ∞[, limk→∞ dp (fk , 0) = 0: Indeed,
Z 1
1
p
lim Np (fk ) = lim χpIk dβ 1 = lim β 1 (Ik ) = lim = 0.
k→∞ k→∞ 0 k→∞ k→∞ k
Given a measure space (X, A, µ), a measurable space (Y, B), and an A-B-measurable
φ : X −→ Y , recall the pushforward measure
µφ : B −→ [0, ∞], µφ (B) = µ φ−1 (B) ,
from Th. 1.69(a).
Theorem 2.45. (a) Let (X, A, µ) be a measure space, (Y, B) a measurable space, φ :
X −→ Y A-B-measurable, and µφ the resulting pushforward measure on (Y, B).
Then, for each f ∈ M+ (B), one has
Z Z
f dµφ = (f ◦ φ) dµ . (2.54)
Y X
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 117
Proof. (a): We note that f ∈ M+ (B) implies f ◦φ ∈ M+ (A) and then use the technique
summarized after the proof of Th. 2.36: First, let f = χB , B ∈ B. Then
Z Z Z
−1
χB dµφ = µφ (B) = µ φ (B) = χφ−1 (B) dµ = (χB ◦ φ) dµ ,
Y X X
such that (2.54) holds. Due to the linearity of the integral, (2.54) then also holds for
each f ∈ S + (B). Now let f ∈ M+ (B) and choose a sequence (uk )k∈N in S + (B) such
that uk ↑ f . Then, clearly, (uk ◦ φ)k∈N is a sequence in S + (A) such that (uk ◦ φ) ↑ (f ◦ φ),
implying
Z Z Z Z
f dµφ = lim uk dµφ = lim (uk ◦ φ) dµ = (f ◦ φ) dµ ,
Y k→∞ Y k→∞ X X
proving (2.54) for each f ∈ M+ (B). The remaining assertion of (a) follows by applying
(2.54) to (Re f )± and (Im f )± .
(b): According to Th. 1.73(a), φ is µn -µn -measurable and (µn )φ = | det φ|−1 µn . Thus,
if f : Rn −→ K̂ is µn -measurable, then so is f ◦ φ. Using (a), if f : Rn −→ K̂ is
nonnegative µn -measurable or µn -integrable, then f ◦ φ has the same property and
Z Z Z
n (2.54) n −1
(f ◦ φ) dµ = f d(µ )φ = | det φ| f dµn ,
Rn Rn Rn
Below, we provide some dense subsets of Lp -spaces (for p < ∞ – L∞ turns out to
have a less benign structure in most situations). Dense subsets are, e.g., useful, since
continuous functions are already uniquely determined by values on dense subsets. Recall
that a topological space is called separable if, and only if, it has a countable dense subset.
Separable spaces are typically much easier to handle than nonseparable spaces. We will
provide some sufficient conditions for Lp -spaces to be separable.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 118
Remark 2.46. Let (X, A, µ) be a measure space, p ∈]0, ∞], and let N denote the set
of measurable f : X −→ K that vanish µ-almost everywhere. Let ι : LpK (µ) −→ LpK (µ),
ι(f ) := [f ], be the quotient map. Consider F ⊆ Lp and F := ι(F) ⊆ Lp . Then F is
dense in Lp if, and only if, F is dense in Lp (in particular, Lp is separable if, and only if,
Lp is separable): Note that dp (f, g) = dp (ι(f ), ι(g)) for each f, g ∈ Lp . Let F be dense
in Lp . Given ι(g) ∈ Lp and ǫ > 0, choose f ∈ F such that dp (f, g) < ǫ. Then ι(f ) ∈ F
and dp (ι(f ), ι(g)) < ǫ, showing F to be dense in Lp . Conversely, let F be dense in Lp .
Given g ∈ Lp and ǫ > 0, choose f ∈ F such that dp (ι(f ), ι(g)) < ǫ. Then dp (f, g) < ǫ,
showing F to be dense in Lp .
i.e. the vector space of simple functions that are 0 outside a set of finite measure,
is dense in Lp : More precisely,
∀p ∀+ ∃ |φ| ≤ |f | ∧ dp (f, φ) < ǫ .
f ∈L ǫ∈R φ∈S0
is dense in Lp .
(c) If there exists a countable semiring S on X such that A = σ(S) and µ↾S is σ-finite,
then Lp is separable.
(d) If there exists a countable semiring S on X such that E = σ(S), E ⊆ A ⊆ Ẽ, where
(X, Ẽ, α̃) is the completion of (X, E, α), α := µ ↾E , µ = α̃ ↾A , and µ ↾S is σ-finite,
then Lp (µ) is separable.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 119
proving the assertion for f ≥ 0. For a general f ∈ Lp , we proceed in the usual way,
letting f1 := (Re f )+ , f2 := (Re f )− , f3 := (Im f )+ , f4 := (Im f )− . Then each fk ∈ Lp
and fk ≥ 0, k = 1, . . . , 4. Given ǫ > 0, choose φk ∈ S0 such that φk ≤ fk and
dp (fk , φk ) < 4ǫ and let φ := φ1 − φ2 + i(φ3 − φ4 ) ∈ S0 . Then
and, for p ≥ 1,
4
X
Np (f − φ) = Np f1 − φ1 − (f2 − φ2 ) + i((f3 − φ3 ) − (f4 − φ4 )) ≤ Np (fk − φk ) < ǫ.
k=1
For p < 1, one replaces Np by Npp in the above estimate, thereby completing the proof
of (a).
(b): According to (a), it suffices to show that, given E ∈ A with µ(E) < ∞ and ǫ > 0,
there exist disjoint sets A1 , . . . , AN ∈ S, N ∈ N, such that µ(Ak ) < ∞ and
! Z
XN p
Npp χE − χAk = χE − χSNk=1 Ak dµ
k=1 X
N
! N
! !
[ [
=µ E\ Ak +µ Ak \E < ǫ.
k=1 k=1
To obtain the Ak , we use the Carathéodory extension Th. 1.38 together with the unique-
ness result of Cor. 1.46(a) (which holds, since µ↾S is σ-finite) to see that µ on A = σ(S)
must be the restriction of the outer measure on P(X) induced by µ on S. Thus, if
R := ρ(S) is the generated ring, then
(∞ ∞
)
(1.30) X [
µ(E) = inf µ(Ak ) : (Ak )k∈N sequence in S, E ⊆ Ak
( k=1
∞
k=1
∞
)
Prop. 1.18(b) X [
= inf µ(Ak ) : (Ak )k∈N sequence in R, E ⊆ Ak
( k=1
∞
k=1
∞
)
X [
= inf µ(Ak ) : (Ak )k∈N disjoint sequence in R, E ⊆ Ak
k=1 k=1
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 120
( ∞ ∞
)
Prop. 1.18(a) X [
= inf µ(Ak ) : (Ak )k∈N disjoint sequence in S, E ⊆ Ak .
k=1 k=1
In consequence,
S sinceSµ(E) < ∞, there exists a sequence of disjoint sets (Ak )k∈N ∈ S,
E⊆ ∞ A
k=1 k and µ ∞ ǫ
k=1 Ak < µ(E) + 2 . Then
N
! N
! ! ∞
!
[ [ [ ǫ
lim µ E \ Ak = 0 and lim µ Ak \ E ≤ µ Ak − µ(E) < .
N →∞
k=1
N →∞
k=1 k=1
2
It turns out that, in many respects, L∞ is typically less benign than Lp for p < ∞. For
example, in general, the space S0 of Th. 2.47(a) is not dense in L∞ and, in the situations
of Th. 2.47(c),(d),(e), L∞ is typically not separable (cf. Rem. 2.55 below).
Definition 2.48. (a) Let (X, T ) be a topological space. Given a function f : X −→ K,
we call the set
supp f := {x ∈ X : f (x) 6= 0}
the support of f . Continuous functions whose support is a compact subset of X are
of particular importance, giving rise to the notation
Cc (X) := Cc (X, K) := f ∈ C(X, K) : supp f is compact .
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 121
where (2.56) is defined to mean that, for each sequence (xk )k∈N in Rn such that
limk→∞ kxk k = ∞, one has limk→∞ f (xk ) = 0). Define
C0 (O) := C0 (O, K) := (f ∈ C(O, K) : f vanishes at infinity .
2
Clearly, Cc (O) ( C0 (O) (if O = Rn , then f (x) := e−kxk2 defines f ∈ C0 (O) \ Cc (O);
2
if O 6= Rn , then f (x) := dist(x, Oc ) · e−kxk2 defines f ∈ C0 (O) \ Cc (O)).
Theorem 2.49. Let n ∈ N and let O ⊆ Rn be open and consider (O, A, µn ), where
(A, µn ) is either (B n , β n ) or (Ln , λn ).
Proof. (a): It suffices to consider (A, µn ) = (Ln , λn ). Due to Th. 2.47(a), it suffices to
show that, for each A ∈ A with λn (A) < ∞ and each ǫ > 0, there exists g ∈ Cc (O) such
that Npp (χA − g) < ǫ. Let δ := min{ 2ǫ , ( 2ǫ )p }. Choose k ∈ N sufficiently large such that,
for B := A ∩ [−k, k]n , one has λn (A) ≤ λn (B) + δ. Then Npp (χA − χB ) < δ. According to
(1.41), there exists a closed set K ⊆ Rn and an open set V ⊆ Rn such that K ⊆ B ⊆ V
and λn (V \ K) < δ. If we let U := V ∩ O ∩ (] − k − 1, k + 1[)n , then U is still open with
B ⊆ U and λn (U \ K) < δ, but U is also bounded and U ⊆ O. Moreover, since K is a
subset of the bounded set B, K is compact. If K = ∅, then λn (B) < δ and λn (A) < ǫ,
such that we can choose g ≡ 0. Thus, consider K 6= ∅. Then, using the continuity of
dist (cf. [Phi16b, Ex. 2.6(b)]), we observe dist(K, U c ) > 0 and define the continuous
function
dist(x, K)
g : O −→ R, g(x) := 1 − min 1, . (2.58)
dist(K, U c )
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 122
we obtain Z Z
Npp (χB − g) = p
|χB − g| dλ ≤ n
|χU − χK |p dλn < δ.
O O
Then, for p < 1, Npp (χA
− g) ≤ − χB ) + Npp (χB − g) < 2ǫ + 2ǫ
Npp (χA = ǫ. For p ≥ 1,
Np (χA − g) ≤ Np (χA − χB ) + Np (χB − g) < 2δ 1/p ≤ 2 2ǫ = ǫ, completing the proof of (a).
(b): Seeking a contradiction,
R assume there p exists ǫ > 0 such that for each s > 0, there
exists s > t > 0 with Rn f(x + t) − f (x) dµn (x) ≥ ǫ. Using (a), choose g ∈ Cc (Rn , K)
R p
such that Rn f (x) − g(x) dµn (x) < 3ǫ . Since g is continuous and, thus, bounded on
the compact set supp g, there exists M ∈ R+ such that 2p |g|p ≤ M . Choosing r > 0
such that supp g ⊆ Br (0) and letting B := B r (0), M χB is integrable and dominates
x 7→ |g(x+t)−g(x)|p for all sufficiently small t ≥ 0. Thus, by the dominated convergence
Th. 2.20,
Z Z
p n Th. 2.20 p g cont.
lim
g(x + t) − g(x) dµ (x) = lim g(x + t) − g(x) dµn (x) = 0.
t→0 Rn Rn t→0
In consequence, Z
∃ ∀ g(x + t) − g(x)p dµn (x) < ǫ
s>0 0≤t<s Rn 3
and, for each 0 ≤ t < s,
Z
f (x + t) − f (x)p dµn (x)
Rn
Z Z
p n
≤ f (x + t) − g(x + t) dµ (x) + g(x + t) − g(x)p dµn (x)
Rn Rn
Z
g(x) − f (x)p dµn (x) < ǫ + ǫ + ǫ = ǫ.
(2.55)
+
Rn 3 3 3
This contradiction proves (2.57).
We will see in Th. 2.54(d) below that even Cc∞ (O, K) is dense in LpK (µn ) for p ∈]0, ∞[.
One can actually extend Th. 2.49(a) to measure spaces (X, A, µ), where µ is a suitable
measure on a locally compact Hausdorff space (see, e.g., [Rud87, Th. 3.14]).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 123
Remark and Definition 2.50. Let n ∈ N and consider (Rn , A, µn ), where (A, µn ) is
either (B n , β n ) or (Ln , λn ). Consider f, g ∈ L1K (µn ). Then
is µ2n -measurable: Indeed, the function A : R2n −→ R2n , A(x, y) := (x − y, y), is linear
and bijective and, thus, µ2n -µ2n -measurable. The projections π1 , π2 : R2n −→ Rn ,
π1 (x, y) = x, π2 (x, y) = y are µ2n -µn -measurable. Thus, as f, g are µn -measurable, and
h = (f ◦ π1 ◦ A) · (g ◦ π2 ), h is µ2n -measurable. Next, we compute
Z Z Z Z
f (x − y)g(y) dµn (x) dµn (y) = g(y) f (x − y) dµn (x) dµn (y)
Rn Rn Rn Z Rn
(2.55)
= kf k1 g(y) dµn (y) = kf k1 kgk1 < ∞.
Rn
Since h is measurable, we may apply Th. 2.36(c) to conclude that h is µ2n -integrable
and there exists a µn -null set N ⊆ Rn such that y 7→ h(x, y) is µn -integrable for each
x ∈ N c . We now define
(R
n Rn
f (x − y)g(y) dµn (y) for x ∈ N c ,
f ∗ g : R −→ K, (f ∗ g)(x) := (2.59)
0 for x ∈ N ,
and call this function the convolution of f and g. From our considerations above, we
then know f ∗ g ∈ L1K (µn ). If f˜, g̃ ∈ L1K (µn ) with f = f˜ µn -a.e. and g = g̃ µn -a.e.,
then, for each x ∈ Rn , we have f (x − y)g(y) = f˜(x − y)g̃(y) for a.e. y ∈ Rn . Thus,
y 7→ f˜(x − y)g̃(y) is µn -integrable for each x ∈ N c , and f˜ ∗ g̃ = f ∗ g. In particular,
convolution “∗” is well-defined as a map from L1K (µn ) × L1K (µn ) to L1K (µn ).
∀ f ∗ g = g ∗ f.
f,g∈L1K (µn )
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 124
∀ (f + g) ∗ h = f ∗ h + g ∗ h µn -a.e.
f,g,h∈L1K (µn )
∀ (f ∗ g) ∗ h = f ∗ (g ∗ h) µn -a.e.
f,g,h∈L1K (µn )
proving associativity.
R
(d): If (f ∗ g)(x) 6= 0, then Rn f (x − y)g(y) dµn (y) 6= 0, i.e. there exists y ∈ supp g such
that x − y ∈ supp f , showing x = y + (x − y) ∈ supp f + supp g and proving (d).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 125
Definition 2.52. Let n ∈ N. We call a sequence (ϕk )k∈N in L1 (β n ) a Dirac sequence if,
and only if, it has the following properties (i) – (iii):
(a) Clearly, (
n (2/k)−n for x ∈ [− k1 , k1 ]n ,
ϕk : R −→ K, ϕk (x) :=
0 otherwise,
defines a (discontinuous) Dirac sequence.
0 otherwise,
where
Z −1
1
ck := exp − dx , B := {x ∈ Rn : kxk2 < 1/k},
B k − kxk22
−2
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 126
defines a Dirac sequence in Cc∞ (Rn ): Clearly, each ϕk maps into R+ 0 , kϕk k1 = 1
by the definition of ck , and supp ϕk = Bk ,k·k2 (0), showing (ϕk )k∈N to be a Dirac
−1
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 127
−1
This, together with (2.60), t−1
γ = (−2xν − tγ ) α(k, xγ ) , and limγ→∞ Qp,j (xγ ) =
Qp,j (x) ∈ R, implies limγ→∞ t−1γ ∂p ϕk (xγ ) = 0 and, thus,
∂p ϕk (xγ ) − 0
∂ν ∂p ϕk (x) = lim = 0,
γ→∞ tγ
(a) If the sequence (ϕk )k∈N in L1 (β n ) is a Dirac sequence according to Def. 2.52, then
(c) If f ∈ L1 (λn ) and g ∈ Cc∞ (Rn ), then f ∗ g ∈ C ∞ (Rn ), and, for each p =
(p1 , . . . , pk ) ∈ {1, . . . , n}k , k ∈ N,
(d) If O ⊆ Rn is open, then Cc∞ (O, K) is dense in LpK (O, Ln , λn ) for each p ∈]0, ∞[.
Since (ϕk )k∈N is a Dirac sequence, there exists N ∈ N such that, for each k > N ,
supp ϕk ⊆ Bδ (0). Thus, for each k > N ,
Z Z
n
kϕk ∗ f − f k1 = ϕ (x − y)f (y) dλ n
(y) − f (x) dλ (x)
k
Rn Rn
Z Z
= ϕk (y) f (x − y) − f (x) dλn (y) dλn (x)
R n Rn
Z Z
Fubini
≤ ϕk (y) f (x − y) − f (x) dλn (x) dλn (y)
Rn Rn
Z
≤ ϕk (y) ǫ dλn (y) = ǫ,
Rn
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 128
proving (a).
(b): Let f, g ∈ L1 (λn ). Since convolution is commutative, it suffices to consider the
case, where g is bounded. If g is bounded, then, for each x ∈ Rn , y 7→ f (y)g(x − y) is
λn -integrable, implying y 7→ f (x − y)g(y) to be λn -integrable as well. If x, h ∈ Rn , then
Z
|(f ∗ g)(x + h) − (f ∗ g)(x)| ≤ f (x + h − y) − f (x − y) g(y) dµn (y)
Rn
Z
= f (y + h) − f (y) g(x − y) dµn (y)
Rn
Z
≤ kgk∞ f (y + h) − f (y) dµn (y) .
Rn
Thus, according to (2.57), for each ǫ > 0, there exists δ > 0 (not depending on x ∈ Rn ),
such that |(f ∗g)(x+h)−(f ∗g)(x)| < ǫ for each h ∈ Bδ (0), proving f ∗g to be uniformly
continuous.
(c): Let f ∈ L1 (λn ), g ∈ Cc∞ (Rn ), j ∈ {1, . . . , n}. Clearly, ∂j g ∈ Cc∞ (Rn ) as well,
and ∂j g is uniformly continuous by [Phi16b, Th. 3.20]. Thus, given ǫ > 0, there exists
δ > 0, such that |∂j g(u) − ∂j g(v)| < ǫ for each u, v ∈ Rn with ku − vk1 < δ. Let ej
denote the jth standard unit vector of Rn . Then, for each x ∈ Rn and each real number
0 6= t ∈] − δ, δ[, one computes
(f ∗ g)(x + tej ) − (f ∗ g)(x)
− f ∗ (∂j g) (x)
t
Z
g(x + te j − y) − g(x − y)
= f (y) − ∂j g(x − y) dλ (y)
n
Rn t
Z Z t
f (y)
= ∂j g(x + sej − y) − ∂j g(x − y) dλ (s) dλ (y)
1 n
Rn t 0
≤ kf k1 ǫ,
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 129
Using the uniform continuity of f , we choose δ > 0 such that |f (u) − f (v)| < ǫ ·
(β n (supp f ) + 1))−1/p for each u, v ∈ Rn with ku − vk < δ. As
\
supp f = Kν , Kν := {x ∈ Rn : dist(x, supp f ) ≤ ν −1 },
ν∈N
we can choose ν0 ∈ N such that β n (Kν0 ) < β n (supp f ) + 1 and ν0−1 < α. Since (ϕk )k∈N is
a Dirac sequence,
R we can choose k0 ∈ N, such that supp ϕk0 ⊆ Bβ (0), β := min{δ, ν0−1 }.
Then, using Rn ϕk0 dλn = 1,
Z
(f ∗ ϕk0 )(x) − f (x) ≤ |f (x − y) − f (x)| ϕk0 (y) dλn (y)
ZR
n
≤ ǫ · (β n (supp f ) + 1)−1/p .
We note, using Prop. 2.51(d), supp(f ∗ ϕk0 ) ⊆ supp f + supp ϕk0 ⊆ Kν0 ⊆ O, since
β ≤ ν0−1 < α. Thus, the estimate
Remark 2.55. As mentioned after the proof of Th. 2.47, L∞ is typically less benign
than Lp for p < ∞:
(a) If µ(X) = ∞, then the space S0 of Th. 2.47(a) is not dense in L∞ : If f ∈ S0 , then
there exists A ∈ A such that µ(A) < ∞ and f ↾Ac = 0, i.e. kf −χX k∞ = kf −1k∞ ≥ 1
(note µ(Ac ) = ∞).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 130
(c) Let O ⊆ Rn be open. In contrast to the case p < ∞ of Th. 2.54(d), Cc (O, K) is
not dense in L∞ n
K (O, A, µ ) (using the same notation as in (b) above): Indeed, the
closure of Cc∞ (O) with respect to k · k∞ turns out to be the space C0 (O) as defined
in Def. 2.48(b) (see [Rud87, Th. 3.17]).
Definition 2.56. Consider (Rn , Ln , λn ), n ∈ N. For each f ∈ L1C (λn ), define
Z
fˆ : R −→ C, fˆ(t) := (2π)
n −n/2
e−i ht,xi f (x) dλn (x) , (2.63a)
Rn
Z
ˇ n ˇ ˆ
f : R −→ C, f (t) := f (−t) = (2π) −n/2
ei ht,xi f (x) dλn (x) , (2.63b)
Rn
where h·, ·i denotes the Euclidean scalar product on Rn (note that, since |e−i ht,xi | =
|ei ht,xi | = 1, the integrability of f implies the integrability of both integrands in (2.63)).
We call fˆ the Fourier transform of f and fˇ the inverse Fourier transform of f (the
name for fˇ is due to Th. 2.59 below). Note that, if f = g λn -a.e., then fˆ = ĝ and fˇ = ǧ,
i.e. fˆ and fˇ are also well-defined for f ∈ L1C (λn ).
Theorem 2.57. Let n ∈ N and let f, g ∈ L1C (λn ).
(a) Both Fourier transform and inverse Fourier transform are linear, i.e. if α, β ∈ C,
then (αf + βg)∧ = αfˆ + βĝ and (αf + βg)∨ = αfˇ + βǧ.
(c) f[
∗ g = (2π)n/2 fˆ · ĝ.
f ∗ : Rn −→ C, f ∗ (x) := f (−x),
fa : Rn −→ C, fa (x) := f (a + x),
(Mr f ) : Rn −→ C, (Mr f )(x) := rn f (rx).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 131
Q
(e) Let p = (p1 , . . . , pk ) ∈ {1, . . . , n}k , k ∈ N. Using the notation xp := kj=1 xpj for
each x ∈ Rn , and assuming ∂p f exists as well as xq f ∈ L1 for each q = (pα , . . . , pk )
with α ∈ {1, . . . , k}, we have
∂p fˆ = (−i)k (xp f )∧ .
Since, for each t ∈ Rn , the integrands in (2.63a) are dominated by the integrable |f |, fˆ
is continuous by Th. 2.22. We note that, for each t, x ∈ Rn , t 6= 0,
π
t, x + t = ht, xi + π,
ktk22
implying (since e−iπ = −1)
Z Z
n/2 ˆ −i ht,xi n (2.55) −i ht,xi π
(2π) f (t) = e f (x) dλ (x) = − e f x+ 2
t dλn (x)
Rn Rn ktk2
and
Z
2 (2π) n/2
|fˆ(t)| ≤ f (x) − f x + π t dλn (x) → 0 for t → ∞
ktk22
Rn
by (2.57).
(c): For each t ∈ Rn , we compute
Z Z
[
f ∗ g(t) = (2π) −n/2
e −i ht,xi
f (y)g(x − y) dλn (y) dλn (x)
Z R Z R
n n
Fubini −n/2 −i ht,x−yi
= (2π) e g(x − y) dλ (x) e−i ht,yi f (y) dλn (y)
n
Rn Rn
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 132
proving (2.64a);
Z
(2.55)
fba (t) = (2π)−n/2 e−i ht,xi f (a + x) dλn (x) = ei ha,ti fˆ(t),
Rn
proving (2.64b);
Z
d t
M r f (t) = (2π)
−n/2
e −i ht,xi
r f (rx) dλ (x) = fˆ
n n
,
Rn r
proving (2.64c);
Z
(e −i ha,xi ∧
f ) (t) = (2π) −n/2
e−i ht,xi e−i ha,xi f (x) dλn (x) = fˆ(t + a) = (fˆ)a (t),
Rn
proving (2.64d).
(e): Let j ∈ {1, . . . , n}. According to the hypothesis, |xj f | is integrable and dominates
the corresponding derivative of the integrand of (2.63a) and we can apply Cor. 2.24 to
obtain
Z
ˆ ˆ
∂j f (t) = ∂tj f (t) = (2π) −n/2
(−i) e−i ht,xi xj f (x) dλn (x) = (−i) (xj f )∧ (t).
Rn
f : R −→ C, f := χ]−a,a[ .
Note that fˆ ∈
/ L1 , since
Z Z ∞ Z ∞
−1/2 −1 ˆ | sin(ta)| | sin x|
(2 (2π) ) |f | ≥ dt = dx
R 0 t 0 x
X∞ Z kπ ∞
1 2 X1
≥ | sin x| dx = = ∞.
k=1
kπ (k−1)π π k=1
k
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 133
(b) Consider
φ : R −→ C, φ(x) := max{0, 1 − |x|}. (2.65a)
1
Then φ ∈ L and we show
2
−1/2 1 − cos t [Phi16a, (I.1c)] −1/2 sin(t/2)
φ̂ : R −→ C, φ̂(t) = (2π) 2 = (2π) .
t2 t/2
(2.65b)
We compute, for each t ∈ R,
Z 1
−1/2
φ̂(t) = (2π) e−itx (1 − |x|) dx .
−1
R ∞|φ̂| is continuous
(the first integral is finite, since on Rthe bounded set [0, 1], the
−1/2 ∞ −2
1
second integral is also finite, as 1 |φ̂| dλ ≤ 4(2π) 1
t dt = 4(2π)−1/2 ). We
now show
(φ̂)∨ = φ : (2.65d)
For each x ∈ R, we obtain
Z 2 Z
∨ 1 sin(t/2)
π (φ̂) (x) = e i tx
dt = t−2 (1 − cos t) cos(tx) dt .
2 R t/2 R
Note
cos(tx + t) + cos(tx − t)
= cos(tx) cos t − sin(tx) sin t + cos(tx) cos t − sin(tx) sin(−t)
= 2 cos(tx) cos t.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 134
Thus,
Z
∨
∞
−2 1
π (φ̂) (x) = 2 t cos(tx) − cos(t(x + 1)) + cos(t(x − 1)) dt
0 2
h i∞
= 2 − t−1 (1 − cos t) cos(tx)
0
Z ∞
−1 1
−2 t x sin(tx) − (x + 1) sin(t(x + 1)) + (x − 1) sin(t(x − 1)) dt .
0 2
We proceed to evaluate each term:
h i∞ ∞
−1 1 − cos t
− t (1 − cos t) cos(tx) = t cos(tx) = 0.
0 t2 0
R∞
For x = 0, we have t−1 x sin(tx) = 0. For x > 0, we have
0
Z ∞ Z ∞
−1 sin u (F.2) π
t x sin(tx) = x du = x .
0 0 u 2
For x < 0, we have
Z ∞ Z −∞
−1 sin u (F.2) π
t x sin(tx) = x du = −x .
0 0 u 2
Altogether, we obtain
∨ π x+1 x−1
∀ π (φ̂) (x) = − −x + + = 0,
x≤−1 2 2 2
∨ x+1 x−1
∀ π (φ̂) (x) = −π −x − + = π (x + 1),
−1≤x≤0 2 2
∨ x+1 x−1
∀ π (φ̂) (x) = −π x − + = π (−x + 1),
0≤x≤1 2 2
∨ π x+1 x−1
∀ π (φ̂) (x) = − x− − = 0,
1≤x 2 2 2
showing (2.65d).
(c) Let n ∈ N. We will use the results of (b) to compute the Fourier transforms of
the members of the Dirac sequence (ϕk )k∈N of Ex. 2.53(b), and we will show that
(2.65d) still holds for φ replaced by ϕk . Recall that
n
Y
n
ϕk : R −→ C, ϕk (x) := max{0, k − k 2 |xj |}.
j=1
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 135
Using φ from (2.65a) and the notation from Th. 2.57(d), we define, for each k ∈ N,
ck = k (M1/k φ̂) and we can apply (2.64c) again to obtain, for each
In other words, φ
x ∈ R,
ck )∨ (x) = (d
(φ
b̂
ck )(−x) = k φ(−kx)
φ
(2.65d)
= k(φ̂)∨ (kx) = kφ(kx) = φk (x),
showing
∀ ck )∨ = φk .
(φ
k∈N
proving
∀ ϕk ) ∨ = ϕk .
(c (2.66b)
k∈N
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 136
Due to Th. 2.57(b), not every function in L1 can be the Fourier transform of an L1
function. On the other hand, we know from Ex. 2.58(a) that the Fourier transform of
an L1 function is not necessarily in L1 . Still one has the following inversion Th. 2.59(a)
for the Fourier transform of L1 functions. But the point is that, due to the mentioned
shortcomings of the Fourier transform for L1 functions, one has to assume fˆ ∈ L1 in Th.
2.59(a). There is a subspace S ( L1 ∩ L2 , the space of so-called Schwartz functions, on
which the Fourier transform is, indeed, bijective, but the details are beyond the scope
of the present class.
Theorem 2.59 (Inversion Theorem). Let n ∈ N.
(b) If f, g ∈ L1C (λn ), then fˆ = ĝ implies f = g λn -a.e. (i.e. the Fourier transform is
injective as a map on L1C (λn )).
Proof. (a): We will make use of the fact that, from (2.66b) of Ex. 2.58(c), we already
know ϕk = (c ϕk )∨ for each member of the Dirac sequence (ϕk )k∈N of Ex. 2.53(b). Since
ck is bounded by Th. 2.57(b) and fˆ ∈ L1 , we have ϕ
ϕ ck fˆ ∈ L1 as well. Thus, we obtain,
n
for each x ∈ R ,
Z
(c ˆ ∨
ϕk f ) (x) = (2π) −n/2
ck (t)fˆ(t) dλn (t)
ei hx,ti ϕ
R n
Z Z
−n/2 −n/2 i hx,ti
= (2π) (2π) e ck (t)
ϕ e−i ht,zi f (z) dλn (z) dλn (t)
ZR R
n n
Z
Fubini −n/2 −n/2 i ht,x−zi
= (2π) (2π) f (z) e ck (t) dλn (t) dλn (z)
ϕ
Rn Rn
Z
= (2π)−n/2 f (z) (c ϕk )∨ (x − z) dλn (z)
ZR
n
(2.66b)
= (2π)−n/2 f (z) ϕk (x − z) dλn (z) = (2π)−n/2 (f ∗ ϕk )(x). (2.67)
Rn
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 137
Recalling that fˇ(t) = fˆ(−t), we can apply Th. 2.57(b) again to conclude
lim
(2π)n/2 (cϕk fˆ)∨ − (fˆ)∨
= lim
(2π)n/2 (c ϕk fˆ)∧ − (fˆ)∧
k→∞ ∞ k→∞ ∞
≤ (2π)−n/2 lim
(2π)n/2 ϕck fˆ − fˆ
= 0,
k→∞ 1
i.e. uniform convergence (2π)n/2 (c ϕk fˆ)∨ → (fˆ)∨ . Applying Prop. G.4 of the Appendix
then yields, for each r ∈ R+ ,
Z Z
ˆ n (G.4a)
∨
|(f ) − f | dλ = lim |(2π)n/2 (cϕk fˆ)∨ − f | dλn
Br (0) k→∞ B (0)
Z r
(2.67)
= lim |(f ∗ ϕk ) − f | dλn
k→∞ Br (0)
(2.61)
≤ lim kϕk ∗ f − f k1 = 0,
k→∞
(a) Fourier transform is an L2 -isometry. More precisely, if f ∈ L1C (λn ) ∩ L2C (λn ), then
fˆ ∈ L2C (λn ) and Z Z
2 n
|f | dλ = |fˆ|2 dλn .
Rn Rn
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 138
We use this equation for g together with the Cauchy-Schwarz inequality (i.e. (2.49) with
p = q = 2) to estimate, for each x, x′ ∈ Rn ,
Z 2
′ 2
|g(x) − g(x )| ≤ ′ n
f (x + y) − f (x + y) |f (y)| dλ (y)
Rn
Z Z
(2.49) 2 n
≤ f (x + y) − f (x + y) dλ (y) ·
′ 2 n
|f (y)| dλ (y) ,
Rn Rn
where the right-hand side converges to 0 for x → x′ (due to Th. 2.49(b) and f ∈ L2 ).
In other words, the function g is continuous and
Once again, we consider the Dirac sequence (ϕk )k∈N of Ex. 2.53(b). We choose N ∈ N
such that supp ϕk ⊆ Bδ (0) for each k ≥ N . Then,
Z
ϕk (−x)=ϕk (x)
∀ |(g ∗ ϕk )(0) − g(0)| = g(x) − g(0) ϕk (x) dλ (x) < ǫ
n
k≥N
Rn
Since g ∗ ϕk ∈ L1 , we may form its Fourier transform and use Th. 2.57(c) to conclude
g\∗ ϕk = (2π)n/2 ĝ · ϕ
ck ∈ L1 (since ĝ is bounded by Th. 2.57(b) and ϕ
ck ∈ L1 (e.g. due to
(2.65c), (2.66a), and Tonelli)). Thus, according to the inversion Th. 2.59(a),
∨ ∨
(2π)n/2 ĝ · ϕ
ck = g\ ∗ ϕk = g ∗ ϕk ,
where the equality even holds exactly (not merely λn -a.e.), since g ∗ ϕk is continuous by
Th. 2.54(b). Next, we apply Fatou (Prop. 2.19) together with limk→∞ ϕ ck (t) = (2π)−n/2
and ĝ = f\ ∗ f ∗ = (2π)n/2 fˆ fˆ = (2π)n/2 |fˆ|2 to obtain
Z Z Z
ˆ 2 n
|f | dλ = (2π) n/2 ˆ 2 n
ck |f | dλ ≤ lim inf
lim ϕ ck ĝ dλn
ϕ
Rn Rn k→∞ k→∞ Rn
Z
n/2 ∨
= (2π) lim inf (cϕk ĝ) (0) = lim inf (g ∗ ϕk )(0) = |f |2 dλn . (2.68)
k→∞ k→∞ Rn
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 139
Here fa , fb ∈ L1 ∩ L2 and, even though fba , fbb are not integrable (according to Ex.
2.58(a)), fba fbb turns out to be integrable.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 140
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 141
Proof. Let ǫ > 0. The key idea is to write I ∈ S as the union of sufficiently small
intervals Iν and, on each Iν , to approximate φ by a suitable affine map xν + Lν . It will
be convenient to use the max-norm k·kmax = k·k∞ on Rn . Recalling Dφ(x), D(φ−1 )(x) ∈
2 2
Rn , it would also be convenient to use the corresponding operator norm k · k∞ on Rn ,
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 142
2
since it satisfies kT (x)k∞ ≤ kT k∞ kxk∞ for each T ∈ Rn and each x ∈ Rn . But
if one is not familiar with the concept of an operator norm, one can avoid it, one
2
can recall from [Phi16b, Lem. 5.4] that there exists a norm k · kHS on Rn , satisfying
2
kT (x)k2 ≤ kT kHS kxk2 for each T ∈ Rn and each x ∈ Rn . Thus, given an arbitrary
2
norm k · k on Rn , by the equivalence of norms, there exists a constant c0 ∈ R+ such
that
∀ 2 ∀ n kT (x)k∞ ≤ c0 kT k kxk∞ .
T ∈Rn x∈R
Due to [Phi16b, Th. 4.38], there exists a constant c1 ∈ R+ such that, given an open
convex set C ⊆ O and α ∈ R+ with
∀ ∀
Dha (x)
≤ α,
a∈I x∈C
On the compact set I ⊆ O, the continuous map x 7→ k(Dφ(x))−1 k attains its max
M := max k(Dφ(x))−1 k : x ∈ I ∈ R+ . (2.75)
Since I ⊆ O, we have dist(I, Oc ) > 0 and we may choose s > 0 such that
Ks := x ∈ Rn : dist(x, I) ≤ s ⊆ O
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 143
Moreover,
(2.73), (2.75)
B ǫ d (0) = Tν Tν−1 (B ǫ d (0)) ⊆ Tν (Bc0 ǫ d (0)),
M M
implying
φ(Iν ) ⊆ φ(aν ) + Tν Iν − aν + Bc0 ǫ d (0) .
Now note that the set Iν − aν + Bc0 ǫ d (0) is contained in a hypercube with side length
d + 2 c0 ǫ d = d(1 + 2 c0 ǫ). In consequence, we obtain
(1.47b)
β n φ(Iν ) ≤ | det Tν | dn (1 + 2 c0 ǫ)n = | det Tν | (1 + 2 c0 ǫ)n β n (Iν )
and
N
X N
X
n n
n
β (φ(I)) = β φ(Iν ) ≤ (1 + 2 c0 ǫ) | det Tν | β n (Iν )
ν=1 ν=1
(2.78)
Z
≤ (1 + 2 c0 ǫ)n | det(Dφ)| dβ n .
I
Since ǫ > 0 was arbitrary, this concludes the proof of Cl. 1. N
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 144
Claim 2. Z
n
∀ n
β φ(A) ≤ | det(Dφ)| dβ n .
A∈BO A
Proof. The statement of Cl. 2 is precisely that (β n )φ−1 ≤ | det(Dφ)|β n holds for the two
n
measures on BO . However, from Cl. 1, we already know that the measures’ respective
restrictions to the generating semiring S satisfy the inequality. Since the measures are
also σ-finite on S, the validity of Cl. 2 is due to Cor. 1.46(b). N
Claim 3. Z Z
n
∀ n)
f dβ ≤ (f ◦ φ)| det(Dφ)| dβ n .
f ∈M+ (U,BU U O
proving Cl. 3 for measurable characteristic functions. It then also holds for each simple
function f ∈ S + (U, BUn ) by the linearity of the integral. For a general f ∈ M+ (U, BUn ), we
choose a sequence (fk )k∈N in S + (U, BUn ) with fk ↑ f and use the monotone convergence
Th. 2.7 to estimate
Z Z Z Z
n n n MCT
f dβ = lim fk dβ ≤ lim (fk ◦ φ)| det(Dφ)| dβ = (f ◦ φ)| det(Dφ)| dβ n
U k→∞ U k→∞ O O
Proof. We apply Cl. 3 with φ−1 instead of φ and (f ◦ φ)| det(Dφ)| instead of f to obtain
Z Z Z
n
−1 −1 n
(f ◦ φ)| det(Dφ)| dβ ≤ (f ◦ φ)| det(Dφ)| ◦ φ | det(Dφ )| dβ = f dβ n ,
O U U
−1
since (Dφ) φ ◦ Dφ−1 = Id by the chain rule. The inequality just proved, together
with the one of Cl. 3, establishes Cl. 4. N
n
Applying Cl. 4 to f = χA with A ∈ BO , we obtain (2.71c) for (A, µn ) = (B n , β n ). In
particular, (2.71c) then says that β n (A) = 0 if, and only if, β n (φ(A)) = 0, showing that
φ provides a bijection between β n -null sets, also implying φ to be λn -measurable. In
consequence, (2.71c) also holds for (A, µn ) = (Ln , λn ). We then obtain (2.71b) (which is
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
2 INTEGRATION 145
equivalent to (2.71a)) by applying (2.71c) with φ−1 instead of φ, completing the proof of
(a). According to Cl. 4, (b) holds for (A, µn ) = (B n , β n ) and f ≥ 0. If f ∈ M+ (U, LnU ),
then, by Prop. G.3, there exists g ∈ M+ (U, BUn ), such that g = f β n -a.e. Then
Z Z Z Z
n
f dλ = g dβ = (g ◦ φ)| det(Dφ)| dβ = (f ◦ φ)| det(Dφ)| dλn ,
n n
U U O O
where the last equality makes, again, use of the fact that φ maps β n -null sets to β n -null
sets. We have, thus, proved (2.72) for f ∈ M+ (U, LnU ). The µn -integrable case then
also follows, in the usual way, by decomposing f into (Re f )± and (Im f )± , completing
the proof of (b). Finally, (c) is obtained by applying (b) with f replaced by g with
(
f (x) for x ∈ φ(A),
g : U −→ K̂, g(x) :=
0 otherwise.
This concludes the proof of the theorem.
Example 2.67. Let
O := R+ ×]0, 2π[, U := R2 \ {(x, 0) : x ≥ 0}. (2.79)
As shown in Sec. H of the Appendix, the function defining polar coordinates, i.e.
φ : O −→ U, φ(r, ϕ) := (r cos ϕ, r sin ϕ), (2.80)
constitutes a C 1 -diffeomorphism with det Dφ(r, ϕ) = r for each (r, ϕ) ∈ O. Since
{(x, 0) : x ≥ 0}, {0} × [0, 2π], R+ + 2
0 × {0}, and R0 × {2π} all are λ -null sets, we can use
+ 2 2
Th. 2.66(b) to conclude, for each f ∈ M (R , L ),
Z Z
2
f dλ = f (r cos ϕ, r sin ϕ) r dλ2 (r, ϕ) . (2.81)
R2 O
proving (2.82).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 146
3.1 Submanifolds of Rn
Roughly speaking, a k-dimensional submanifold of Rn is a subset that, locally, looks
like an open subset of Rk . There are a number of equivalent ways of making this idea
precise, see Def. 3.1 and Th. 3.4 below.
(a) Perhaps the most simple submanifolds of Rn are its affine subspaces: Let k ∈ N,
k < n, and let V ⊆ Rn be a k-dimensional vector space over R. Moreover, let
xM ∈ Rn and M := V + xM . Then M is a k-dimensional submanifold of Rn of class
C ∞ : Let {v1 , . . . , vn } be a basis of Rn such that {v1 , . . . , vk } is a basis of V . Let
A : Rn −→ Rn−k be the linear map defined by
(
0 for i ∈ {1, . . . , k},
A(vi ) :=
vi for i ∈ {k + 1, . . . , n},
and
f : Rn −→ Rn−k , f (x) := A(x − xM ).
We can now let, for each a ∈ M , Oa := Rn , f a := f . We verify Def. 3.1(a),(b) to be
satisfied: First note that f is C ∞ , since f is the composition of a translation with
a linear map, f = A ◦ T−xM . Moreover, for each x ∈ Rn ,
f (x) = A(x − xM ) = 0 ⇔ x − xM ∈ V ⇔ x ∈ V + xM = M,
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 147
rk Df (x) = rk A = n − k,
M = {x ∈ Rn : f (x) = 0},
∀ rk Dφ(x) = k (3.1)
x∈O
(i.e., for each x ∈ O, the k vectors ∂1 φ(x), . . . , ∂k φ(x) are linearly independent
elements of Rn ).
the map that reorders the variables according to π, is a linear isomorphism (in
particular, a C ∞ -diffeomorphism).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 148
αa ∈ Aa ⊆ Rk , β a ∈ Ba ⊆ Rn−k ,
F a (M ∩ Va ) = Ek ∩ Ua , (3.4)
where
Ek := {x ∈ Rn : xk+1 = · · · = xn = 0}. (3.5)
(iv) For each a ∈ M , there exists an M -open neighborhood Ma ⊆ M (open with re-
spect to the relative topology on M ), an open set Wa ⊆ Rk , and a C α -immersion
φa : Wa −→ Rn , such that Ma = φa (Wa ) and φa : Wa −→ Ma is a homeomor-
phism. The maps φa are called (local) charts or (local) parametrizations or (local)
coordinates of M .
Proof. (i) ⇒ (ii): Given a ∈ M , let f a : Oa −→ Rn−k be according to Def. 3.1. Since
rk Df a (a) = n − k, there exists π a ∈ Sn such that
∂πa (k+1) f1a · · · ∂πa (n) f1a
.. ..
det . . (a) 6= 0.
a a
∂πa (k+1) fn−k · · · ∂πa (n) fn−k
Let Ga := Pπa (Oa ). Then Ga ⊆ Rn is open by [Phi16b, Cor. 4.51]. Let τ a := (π a )−1 .
Then Pτ a (Ga ) = Oa , Pτ a (αa , β a ) = a. If we let
ha : Ga −→ Rn−k , ha (x, y) := f a Pτ a (x, y) , x ∈ Rk , y ∈ Rn−k ,
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 149
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 150
proving (iii).
(iii) ⇒ (i): For a ∈ M , let F a : Va −→ Ua be given by (iii) such that (3.4) holds. We
set Oa := Va and define
f a (x) = 0 ⇔ Fk+1
a
(x) = · · · = Fna (x) = 0 ⇔ F (x) ∈ Ek ∩ Ua ⇔ x ∈ M ∩ Oa ,
According to the inverse function theorem [Phi16b, Th. 4.50] there exist open sets
Aa , Ba ⊆ Rk such that c ∈ Aa ⊆ Wa and
ϕ̃a : Aa −→ Ba , ϕ̃a (x) := φaπa (1) (x), . . . , φaπa (k) (x) ,
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 151
ϕa : Ua −→ Ba × Rn−k ,
(
φaπa (j) (x) for 1 ≤ j ≤ k,
ϕaj (x, y) := a
φπa (j) (x) + yi for j = k + i, 1 ≤ i ≤ n − k,
τ a := (π a )−1 , and
Ga : Ua −→ Rn , Gaj (z) := ϕaτa (j) (z).
Then Ga is C α , since φa is C α . Moreover, det DGa = det Dϕ̃a 6= 0, since
Dϕ̃a 0
DG = Pτ a
a
.
Dx (ϕak+1 , . . . , ϕan ) Idn−k
Ga (Ek ∩ Ua ) = M ∩ Va .
Theorem 3.5 (Chart Transition). Let k, n ∈ N with k < n, α ∈ N ∪ {∞}, and let M ⊆
Rn be a k-dimensional submanifold of class C α . For j ∈ {1, 2}, let φj : Wj −→ Mj ⊆ M
be local charts according to Th. 3.4(iv), i.e. Wj ⊆ Rk open, the Mj are M -open, and
both φj are C α -immersions as well as homeomorphisms. Suppose V := M1 ∩ M2 6= ∅.
Then Oj := φ−1j (V ) ⊆ Wj are open and
τ := φ−1
2 ◦ φ1 : O1 −→ O2 (3.7)
is a C α -diffeomorphism. The map τ is then also called a transition between the charts
φ1 and φ2 (note φ1 = φ2 ◦ τ on O1 ).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 152
Proof. Since V is M -open and φj are continuous, both Oj are open. Clearly, τ is bijective
(injective, since the φj are injective, surjective by the choice of the Oj ). Since φ−1
2 is not
a C α -map defined on an open subset of Rn , we can not directly apply the chain rule,
but have to work slightly harder. Fix an arbitrary c1 ∈ O1 and set
a := φ1 (c1 ), c2 := φ−1
2 (a) = τ (c1 ).
F a (M ∩ Va ) = Ek ∩ Ua .
Now we are in a situation, where the chain rule applies (to g and to h), yielding
where P denotes the projection from Rn to Rk . Since rk Dφj (x) = k, Dφj must be
injective. Since F a is a diffeomorphism, D(F a ◦ φj ) must still be injective. Since φj
maps into M ∩ Va and F a maps M ∩ Va into Ek (i.e. the last n − k components of F a ◦ φj
are identically 0 on W̃1 ), both Dg(x) and Dh(x) must be surjective as well as injective,
showing g : W̃1 −→ g(W̃1 ) and h : W̃2 −→ g(W̃2 ) to be C α -diffeomorphisms. Thus,
τ ↾W̃1 = (φ−1
2 ◦ φ1 )↾W̃1 = h
−1
◦g
(b) M ∈ B n , i.e. M is β n -measurable (one can also show that M is even a β n -null set).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 153
Proof. (a): Recall that each open set O ⊆ Rn is the countable union of compact intervals,
n
for example of intervals from the set C := Ic,Q (defined in Ex. 1.8(c)). Let a ∈ M and let
Oa , f a ∈ C 1 (Oa , Rn−kS) be according to Def. 3.1. Then, there exists a sequence (Iia )i∈N
in C such that Oa = i∈N Iia . Thus,
[
M ∩ Oa = M ∩ Oa ∩ Iia .
i∈N
Since f a is continuous, (f a )−1 ({0}) is closed. As Iiai is compact, (f a )−1 ({0}) ∩ Iiai is
compact as well, proving (a).
(b) follows from (a), as each compact set is in B n .
is called the chart’s Gram matrix or its metric tensor. If we denote the columns of Dφ
by ∂i φ, then we can write the components gij of G as follows:
n
X
∀ ∀ gij (x) = h∂i φ(x), ∂j φ(x)i = ∂i φν (x) ∂j φν (x). (3.9)
x∈W i,j∈{1,...,k}
ν=1
The map
γ := γ(φ) : W −→ R, γ(x) := det G(x), (3.10)
is called the chart’s Gram determinant.
Theorem 3.8. Let k, n ∈ N with k < n and let M ⊆ Rn be a k-dimensional submanifold
of class C 1 .
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 154
Proof. (a): Let x ∈ W and set A := Dφ(x). Then G(x) = At A, i.e. G(x)t = (At A)t =
At A = G(x), showing G(x) to be symmetric. If 0 6= y ∈ Rk , then y t G(x)y = y t At Ay =
kAyk2 > 0, since y 6= 0 and A injective (since rk A = k) imply Ay 6= 0. Thus, G(x) is
positive definite. In particular, γ(x) = det G(x) > 0.
(b): For each x ∈ O1 , we have, by the chain rule,
Dφ1 (x) = Dφ2 (τ (x)) Dτ (x).
Thus,
t t
G(φ1 )(x) = Dφ2 (τ (x)) Dτ (x) Dφ2 (τ (x)) Dτ (x) = Dτ (x) G(φ2 )(τ (x))Dτ (x),
implying (3.11).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 155
We call the partition of unity measurable if, and only if, each function ηi ◦ φi is
λk -measurable.
where γ(φi ) denotes the Gram determinant of φi (recall that γ(φi ) is R+ -valued).
If A ⊆ M and χA is σ-measurable, then we call A σ-measurable and define
Z
σ(A) := χA dσ . (3.13)
M
R
We call f : M −→ K̂ σ-integrable if, and only if, f is σ-measurable and M g dσ <
∞ for each g ∈ {(Re f )+ , (Re f )− , (Im f )+ , (Im f − )}. In that case, we define
Z Z Z Z Z
−
f dσ := +
(Re f ) dσ − (Re f ) dσ + i +
(Im f ) dσ − i (Im f )− dσ .
M M M M M
(3.14)
Example 3.10. Let k, n ∈ N with k < n and let M ⊆ Rn be a k-dimensional subman-
ifold of class C 1 .
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 156
(c) Graphs of C 1 -functions are submanifolds coverable by a single local chart: Let
A ⊆ Rk be open and let g : A −→ Rn−k be C 1 . Then the k-dimensional submanifold
and one can show (exercise) that, in this case, the Gram determinant is
n−1
X
+
γ(φ) : A −→ R , γ(φ)(x) = 1 + k ∇ g(x)k22 =1+ |∂i g(x)|2 . (3.16)
i=1
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 157
(a) Given a finite cover C := (φ1 , . . . , φN ), N ∈ N, of local charts for M , there exists a
measurable partition of unity subordinate to C.
(b) The measurability of f : M −→ K̂, as defined in Def. 3.9(c), does not depend on the
chosen local charts covering M . The integral defined in Def. 3.9(c) is well-defined,
i.e.
R the integrals on the right-hand side of (3.12) all make sense and the value of
M
f dσ does neither depend on the choice of covering local charts nor on the choice
of partition of unity. In consequence, the integral defined in (3.14) is well-defined
as well.
S
Then Def. 3.9(b)(i)–(iii) are, clearly, satisfied ((iii), since M = Ni=1 Mi ). It remains
to show that the ηi are measurable, i.e. that the Bi are measurable. As the Mi are
M -open, there exists Oi ⊆ Rn open such that Mi = M ∩ Oi . Since M ∈ B n by Th.
3.6(b), Mi ∈ B n , showing Bi ∈ B n as well.
(b): If f : M −→ [0, ∞], then all integrands on the right-hand side of (3.12) are
nonnegative. If f is σ-measurable, then the integrands are λk -measurable p as they are
the product of the λk -measurable functions f ◦ φi , ηi ◦ φi , and x 7→ γ(φi )(x), the
last function even being continuous. It remains to show the independence of the chosen
charts and partition of unity. Thus, in addition to the chart cover C = (φ1 , . . . , φN ),
φi : Wi −→ Mi and the ηi , let D := (ψ1 , . . . , ψM ), M ∈ N, also be a finite cover of M
via local charts, ψi : Ui −→ Ni , where (α1 , . . . , αM ) constitutes a measurable partition
of unity subordinate to D. Let Vij := Mi ∩ Nj and K := {(i, j) : Vij 6= ∅}. According to
Th. 3.5, for each (i, j) ∈ K, there exists a C 1 -diffeomorphism τij : Uij −→ Wij such that
ψj = φi ◦τij on Uij , where Uij := ψj−1 (Vij ), Wij := φ−1
j (Vij ). Then, for each f : M −→ K̂
and j ∈ {1, . . . , M },
N
X X
f ◦ ψj = (ηi f ) ◦ ψj = (ηi f ) ◦ φi ◦ τij .
i=1 i: (i,j)∈K
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 158
(3.11) X Z q
= ((ηi αj f ) ◦ φi ◦ τij ) | det Dτij | γ(φi ) ◦ τij dλk
(i,j)∈K Uij
(2.72) X Z p
= ((ηi αj f ) ◦ φi ) γ(φi ) dλk
(i,j)∈K Wij
M Z
N X
X N Z
X
p k
p
= ((ηi αj f ) ◦ φi ) γ(φi ) dλ = ((ηi f ) ◦ φi ) γ(φi ) dλk ,
i=1 j=1 Wi i=1 Wi
N
\
Aσ = Bi , Bi := {B ∈ M : φ−1 k
i (B) ∈ L }. (3.18)
i=1
implying
∞
X
σ(A) = σ(Aj ),
j=1
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 159
f σ-measurable ⇔ ∀ f ◦ φi λk -measurable
i∈{1,...,N }
It is then immediate from Def. 3.9(c) that the respective notions of integrability are also
identical. It remains to show that the integral defined in Def. 3.9(c) is the usual integral
with respect to the measure σ. To this end, for the remainder of the proof, denote the
measure σ by λk (to notationally distinguish the two integrals – the notation is also
reasonable as one may interpret the measure σ as a k-dimensional Lebesgue measure
on M ). Let f ∈ S + (Aσ ) be a simple
Pm function. Then there exist A1 , . . . , Am ∈ Aσ and
+
s1 , . . . , sm ∈ R0 such that f = j=1 sj χAj . Thus,
Z N Z
X p N X
X m Z p
k
f dσ = ((ηi f ) ◦ φi ) γ(φi ) dλ = sj ((ηi χAj ) ◦ φi ) γ(φi ) dλk
M i=1 Wi i=1 j=1 Wi
Xm Z m
X Xm Z
(3.13) k
= sj χAj dσ = sj σ(Aj ) = sj λ (Aj ) = f dλk .
j=1 M j=1 j=1 M
Now let f ∈ M+ (Aσ ) and let (fj )j∈N be a sequence in S + (Aσ ) such that fj ↑ f . Then,
using the monotone convergence Th. 2.7,
Z N Z
X p N Z
X p
k MCT
f dσ = ((ηi f ) ◦ φi ) γ(φi ) dλ = lim ((ηi fj ) ◦ φi ) γ(φi ) dλk
M Wi j→∞ Wi
i=1 i=1
Z Z
= lim fj dλk = f dλk .
j→∞ M M
R R
Finally, M f dσ = M f dλk for integrable f : M −→ K̂ follows in the usual way by
decomposing f into (Re f )± , (Im f )± .
Remark
R 3.12. The main benefit of Th. 3.11(c) is that we have identified the integrals
M
f dσ as measure-theoretic integrals in the sense of Sec. 2. In consequence all the
abstract results of Sec. 2 regarding such integrals must hold and are available for integrals
over submanifolds.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 160
Example 3.13. (a) As in Ex. 3.10(b), let I ⊆ R be a nonempty open interval and
let φ : I −→ Rn , n ≥ 2, be C 1 with Dφ(x) 6= 0 for each x ∈ I and such that
φ : I −→ φ(I) is a homeomorphism. We know from Ex. 3.10(b) that M := φ(I) is
a 1-dimensional submanifold covered by the single chart φ and that
2
γ(φ) : I −→ R+ , γ(φ)(x) =
φ′ (x)
2 .
In particular, Z Z
′
l(φ) := σ(M ) = 1 dσ =
φ (x)
dλ1 (x) (3.20)
2
M I
is called the arc length of φ (or M ). Moreover, φ and M are called rectifiable if,
and only if, l(φ) < ∞. We also extend these notions to C 1 -paths φ : I −→ Rn
on general (not necessarily open) intervals by setting l(φ) := l(φ ↾I ◦ ) (where I ◦
denotes the interval’s interior). It is then clear from (3.20) that C 1 -paths defined
on compact intervals are always rectifiable.
(b) Let x0 ∈ [0, 2π] and consider
φ : [0, x0 ] −→ R2 , φ(x) := (cos x, sin x),
φ′ : [0, x0 ] −→ R2 , φ′ (x) = (− sin x, cos x) 6= (0, 0).
p
Then kφ′ (x)k2 = (sin x)2 + (cos x)2 = 1, implying
Z Z x0
′
l(φ) =
1
φ (x) 2 dλ (x) = 1 dx = x0 .
[0,x0 ] 0
This finally proves that the definition of sine and cosine as given in [Phi16a, Def.
and Rem. 8.21] is the same as the high school definition, where cos x and sin x are
defined to be the coordinates of the point p = (p1 , p2 ) ∈ R2 on the unit circle, such
that x is the angle measured in radian (i.e. the arc length of the corresponding circle
section as computed above) between the line segment between (0, 0) and (1, 0) and
the line segment between (0, 0) and p (cf. remarks at the beginning of [Phi16a, Sec.
8.4]).
Example 3.14. (a) Let n ∈ N, n ≥ 2. As in Ex. 3.10(c) (with k := n − 1) let
A ⊆ Rn−1 be open and let g : A −→ R be C 1 . We know from Ex. 3.10(c) that the
(n − 1)-dimensional submanifold
M := {(x, y) ∈ A × R : y = g(x)}
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 161
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 162
showing
∀ γ(Tr ◦ φν )(x) = r2k γ(φν )(x).
x∈Wν
proving (3.28). The integrable case now follows in the usual way by decomposition of f
into (Re f )± and (Im f )± .
Theorem 3.16. Let n ∈ N, n ≥ 2. Let f : Rn −→ K̂ be λn -integrable. Then, for
λ1 -a.e. r ∈ R+ , f is σ-integrable over the sphere Sr (0) = {x ∈ Rn : kxk2 = r} and
Z Z ∞Z Z ∞ Z
n 1 (3.28)
f dλ = f (x) dσ (x) dλ (r) = f (rt) dσ (t) rn−1 dλ1 (r).
Rn 0 Sr (0) 0 S1 (0)
(3.29)
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 163
This formula is sometimes called the coarea formula. It is actually a special case of the
more general coarea formula of geometric measure theory.
Proof. Denote the (open) upper and lower half spaces by H± := {x ∈ Rn : ±xn > 0}
and U := {x ∈ Rn−1 : kxk2 < 1}. We will apply the change of variables given by the
C 1 -diffeomorphism
q
+
φ : U × R −→ H+ , 2
φ(x1 , . . . , xn−1 , r) := rx1 , . . . , rxn−1 , r 1 − kxk2 ,
−1 + −1 x1 xn−1
φ : H+ −→ U × R , φ (x1 , . . . , xn ) := , ..., , kxk2 ,
kxk2 kxk2
Dφ : U × R+ −→ H+ ,
r 0 ··· 0 x1
0 r ··· 0 x2
.. ... ..
Dφ(x1 , . . . , xn−1 , r) :=
. . .
0 ··· 0 r xn−1
p
− √ rx1 − √ rx2 · · · − √rxn−1 1 − kxk22
1−kxk22 1−kxk22 1−kxk22
We compute det Dφ(x, r) via eliminating the nondiagonal elements of the last row by
successively adding, for i = 1, . . . , n − 1, the i-th row multiplied by − √ xi 2 to the
1−kxk2
last row, ending up with an upper triangular matrix, where the upper n − 1 diagonal
elements are r and the last diagonal element is
q n−1
X q
x2i kxk22 1
1 − kxk22 − p = 1 − kxk22 − p = p .
i=1 1 − kxk22 1 − kxk22 1 − kxk22
Thus,
rn−1
det Dφ : U × R+ −→ R+ , det Dφ(x, r) = p .
1 − kxk22
We now apply the change of variables formula (2.72) followed by Fubini to the integrable
function f to obtain
Z Z ∞Z q
n rn−1
f dλ = 2
f rx, r 1 − kxk2 p 2
dλn−1 (x) dλ1 (r)
H+ 0 U 1 − kxk2
Z ∞Z
(3.26)
= f dσ dλ1 (r). (3.30)
0 H+ ∩Sr (0)
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 164
As f is integrable, Fubini also yields that the inner integral must be finite λ1 -a.e. as
claimed. Next, we note that the result of (3.30) still holds with H+ replaced by H− : For
H− , one merely puts a negative sign in front of the last component of φ, then | det Dφ|
remains the same as before and in the middle term of (3.30), one has a “−” in the
second argument of f , which is consistent with what one obtains if one does Ex. 3.14(b)
for the lower hemisphere (i.e. with g replaced by −g). Since Sr (0) ∩ {x ∈ Rn : xn = 0}
is a σ-null set (exercise), the proof is complete.
n n n π n/2
β (Br (0)) = ωn r , ωn = β (B1 (0)) = n .
Γ( 2 + 1)
implying
n π n/2 2 π n/2
σn := σ(S1 (0)) = n ωn = = . (3.31)
Γ( n2 + 1) Γ( n2 )
An application of Th. 3.15 now yields the formula for σ(Sr (0)) = σ(Tr (S1 (0))):
Z Z
(3.28)
σ(Sr (0)) = 1 dσ = rn−1 dσ = σn rn−1 . (3.32)
Tr (S1 (0)) S1 (0)
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 165
(a) A vector τ ∈ Rn is called a tangent vector to M at a if, and only if, there exists
ǫ ∈ R+ and a C 1 -function
ψ : ] − ǫ, ǫ[−→ M,
such that
ψ(0) = a and ψ ′ (0) = τ. (3.34)
The set (we will see in the next theorem it is, indeed, a k-dimensional vector space
over R)
Ta M := {τ ∈ Rn : τ is tangent vector to M at a} (3.35)
is called the tangent space to M at a.
(b) A vector ν ∈ Rn is called a normal vector to M at a if, and only if, it is perpendicular
to all tangent vectors at a (with respect to the Euclidean scalar product), i.e. if,
and only if,
∀ hτ, νi = 0. (3.36)
τ ∈Ta M
We also define
is the space perpendicular to V . We then know from Linear Algebra that V ⊥ is a vector
subspace of Rn , Rn = V + V ⊥ , V ∩ V ⊥ = {0}, and, in particular, dim V ⊥ = n − k.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 166
forms a basis of Na M .
(c) Ta M ⊥ Na M , i.e. if τ ∈ Ta M and ν ∈ Na M , then hτ, νi = 0.
Proof. Set
V1 := span ∂1 φ(c), . . . , ∂k φ(c) ,
U := span ∇ f1 (a), . . . , ∇ fn−k (a) ,
V2 := U ⊥ .
Then V1 , U, V2 all are vector subspaces of Rn . Since ∂1 φ(c), . . . , ∂k φ(c) are linearly in-
dependent, we have dim V1 = k. Since ∇ f1 (a), . . . , ∇ fn−k (a) are linearly independent,
we have dim U = n − k. Then we know from Def. and Rem. 3.20 that dim V2 = k. We
will show
V1 ⊆ Ta M ⊆ V2 . (3.39)
Then dim V1 = dim V2 implies V1 = V2 and V1 = Ta M = V2 , proving (a). Then
Na M = V2⊥ = (Ta M )⊥ , proving (b) and (c). It remains to show (3.39).
“V1 ⊆ Ta M ”: Let τ ∈ V1 . Then
k
X
∃ τ= αi ∂i φ(c).
α1 ,...,αk ∈R
i=1
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 167
showing τ ∈ Ta M .
“Ta M ⊆ V2 ”: Let τ ∈ Ta M . Then there exists ǫ > 0 and a C 1 -function ψ : ] − ǫ, ǫ[−→ M
such that a = ψ(0) and τ = ψ ′ (0). Since a ∈ O, O open, we may actually assume that
ψ maps into M ∩ O = f −1 ({0}). Thus,
∀ fi ◦ ψ ≡ 0,
i∈{1,...,n−k}
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 168
Proof. “⊆ of (3.40)”: Recall Ω = Ω∪∂Ω, where the union is disjoint, as Ω is open. Thus,
according to Def. 3.22(a), we have to show that f a (x) < 0 implies x ∈ / ∂Ω. However,
if x ∈ Ω ∩ Oa with f a (x) < 0, then, by the continuity of f a , there exists an open set
U ⊆ Oa such that f a < 0 on U , showing x ∈/ ∂Ω.
“⊇ of (3.40)”: Let x ∈ Oa with f a (x) = 0. We need to verify x ∈ M . Consider the
auxiliary function
ψ : ] − ǫ, ǫ[−→ R, ψ(s) := f a x + s ∇ f a (x) ,
(i) kν(a)k2 = 1.
∀ a + s ν(a) ∈
/ Ω.
s∈]0,ǫ[
The vector ν(a) is called the outer unit normal to ∂Ω at a. Moreover, the function
ν : ∂Ω −→ Rn , a 7→ ν(a),
is continuous.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 169
Then kν(a)k2 is clear, condition (ii) holds according to (3.41), and ν(a) ∈ Na ∂Ω by Th.
3.21(b). To show ν as a function on ∂Ω is continuous, it suffices to show ν is continuous
at each a ∈ ∂Ω. aHowever, given a ∈ ∂Ω, on Oa ∩ ∂Ω, ν is given by the continuous
function x 7→ k ∇∇ffa (x)k
(x)
2
.
Example 3.25. Let n ∈ N, n ≥ 2. For xM ∈ Rn and r > 0, consider the Euclidean ball
Ω := Br (xM ). Then ∂Ω = {x ∈ Rn : kx − xM k2 = r} and, according to Ex. 3.2(b), for
each a ∈ ∂Ω, we may choose f a := f with
n
X
f : Rn −→ R, f (x) := kx − xM k22 − r2 = (xi − xM,i )2 − r2 ,
i=1
n n
∇ f : R −→ R , ∇ f (x) = 2 x1 − xM,1 , . . . , xn − xM,n .
Thus,
∇ f (a) a − xM
∀ ν(a) = = . (3.43)
a∈∂Ω k ∇ f (a)k2 r
Proposition 3.26. Let n ∈ N, n ≥ 2, and let Ω ⊆ Rn be open, having a C 1 -boundary.
Then, for each a ∈ M := ∂Ω, there exists a permutation π a ∈ Sn and open intervals
Aa ⊆ Rn−1 and Ia ⊆ R, such that
αa ∈ Aa ⊆ Rn−1 , β a ∈ Ia ⊆ R,
where αa := (aπa (1) , . . . , aπa (n−1) ), β a := aπa (n) , and there exists a C 1 -map g a : Aa −→
Ia , satisfying
M ∩ Pτ a (Aa × Ia ) = Pτ a (x, y) ∈ Aa × Ia : y = g a (x) , τ a := (π a )−1 , (3.44a)
and either
Ω ∩ Pτ a (Aa × Ia ) = Pτ a (x, y) ∈ Aa × Ia : y < g a (x) (3.44b)
or
Ω ∩ Pτ a (Aa × Ia ) = Pτ a (x, y) ∈ Aa × Ia : y > g a (x) . (3.44c)
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 170
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 171
(this might not be the same f a as before, but it also satisfies Def. 3.22). We then obtain
∇ f a (a) − ∇ g(αa ), 1 Pπa
ν(a) = a
= p , (αa , β a ) := Pπa (a),
k ∇ f (a)k2 a
1 + k ∇ g(α )k2 2
proving (3.45). For (3.44c), one uses f a (z) := xn + g(x1 , . . . , xn−1 ), x := Pπa (z), again
proving (3.45).
In preparation for the proof of the Gauss-Green Th. 3.30, we prove a special case in the
following lemma:
Lemma 3.28. Let n ∈ N, n ≥ 2. Let A ⊆ Rn−1 be open and let I :=]α, β[ with α, β ∈ R,
α < β, be an open interval. Given a C 1 -function
g : A −→ I,
set
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 172
where
− ∇ g(x̃), 1
ν : M −→ Rn , ν(x̃, xn ) := p . (3.48)
1 + k ∇ g(x̃)k22
Moreover, (3.47) still holds if one has “>” instead of “<” in the definition of Ω and a
“−” in front of the right-hand side of (3.48).
Then
Z xn
∀ ∂n F (x̃, xn ) = f (x̃, xn ), ∂i F (x̃, xn ) = ∂i f (x̃, s) ds
(x̃,xn )∈A×I α
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 173
and, thus,
Z Z Z g(x̃) Z
∂n f dλn = ∂n f (x̃, s) ds dλ n−1
(x̃) = f (x̃, g(x̃)) dλn−1 (x̃)
Ω A
Zα A
(3.48), (3.21)
= = f νi dσ ,
M
thereby completing the proof of the first case. One now obtains the second case by
applying the first to −Ω, −M , −g : −A −→ −I, followed by a change of variables
x 7→ −x.
As a last preparation for the proof of the Gauss-Green Th. 3.30, we construct a C ∞
partition of unity. Here, the partition of unity is subordinate to an open cover of Rn ,
not to an open cover of a submanifold. For ǫ > 0, we consider the cover given by
(B2ǫ (ǫ p))p∈Zn ,
where the balls are defined with respect to the ∞-norm on Rn (i.e. they are open intervals
of side length 4ǫ).
Proposition 3.29. Let n ∈ N, ǫ ∈ R+ . Then there exists a family (ηp )p∈Zn of functions
in Cc∞ (Rn , R), satisfying the following conditions (i) – (iii) (the balls are defined with
respect to the ∞-norm):
Since the support of t 7→ ϕ(t − ǫ p) is [ǫ p − ǫ, ǫ p + ǫ], for each t ∈ R, at least one and at
most two summands in (3.51) are nonzero. In particular, G is well-defined as a function
into R+ . Thus,
ϕ(t)
g : R −→ [0, 1], g(t) := ,
G(t)
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 174
is well-defined, g ∈ Cc∞ (R, R), supp g = [−ǫ, ǫ]. We claim that letting
∀ ηp : R −→ [0, 1], ηp (t) := g(t − ǫ p),
p∈Z
proves the proposition for n = 1: Clearly, (i) and (ii) are satisfied. Since, for each t ∈ R,
X X ϕ(t − ǫ p) X ϕ(t − ǫ p)
ηp (t) = = = 1,
p∈Z p∈Z
G(t − ǫ p) p∈Z
G(t)
As before, ηp ∈ Cc∞ (Rn , R) and (i) and (ii) are immediate consequences of the properties
of g. To obtain (iii), we carry out and induction, assuming the case n − 1 by induction
hypothesis, obtaining, for each x ∈ Rn ,
X X X n−1
Y
ηp (x) = g(xn − ǫ k) g(xi − ǫ qi )
p∈Zn k∈Z q∈Zn−1 i=1
X X n−1
Y
= g(xn − ǫ k) g(xi − ǫ qi )
k∈Z q∈Zn−1 i=1
X
= g(xn − ǫ k) = 1,
k∈Z
The following version of the Gauss-Green theorem is certainly not the most general. For
example, it still holds if, instead of a C 1 -boundary, one just has a so-called Lipschitz
boundary (which, e.g., allows corners if they are sufficiently benign, including polyhedral
sets), see [Kön64].
Theorem 3.30 (Gauss-Green). Let n ∈ N, n ≥ 2. Let Ω ⊆ Rn be open and bounded,
having a C 1 -boundary (then Ω and ∂Ω are compact; in particular, ∂Ω is a finitely cov-
erable (n − 1)-dimensional submanifold of class C 1 ). If Ω ⊆ U ⊆ Rn with U open and
F ∈ C 1 (U, Rn ) (in this context, F is often called a vector field), then
Z Z
n
div F dλ = hF, νi dσ (3.52)
Ω ∂Ω
n
Pn before, ν : ∂Ω −→ R denotes the outer unit normal to Ω, also recall that div F =
(as
i=1 ∂i Fi ).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 175
If supp ηp ⊆ Pτ a (Aa × Ia ), then we apply Lem. 3.28 to obtain, for each i ∈ {1, . . . , n},
Z Z Z
n n
∂i (ηp Fi ) dλ = ∂i (ηp Fi ) dλ = ∂i (ηp Fi ) ◦ Pτ a dλn
Ω
ZΩ∩Pτ (Aa ×Ia ) (Pπa Ω)∩(Aa ×Ia )
a
(3.47)
= (ηp Fi ) ◦ Pτ a νi dσ
(Pπa ∂Ω)∩(Aa ×Ia )
Z Z
(3.45)
= ηp Fi νi dσ = ηp Fi νi dσ .
∂Ω∩Pτ a (Aa ×Ia ) ∂Ω
Now let
Z := {p ∈ Zn : supp ηp ∩ Ω 6= ∅}.
Then
Z XZ XZ XZ
n n n
div F dλ = div(ηp F ) dλ = div(ηp F ) dλ = hηp F, νi dσ
Ω p∈Zn Ω p∈Z Ω p∈Z ∂Ω
XZ Z
= hηp F, νi dσ = hF, νi dσ ,
p∈Zn ∂Ω ∂Ω
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
3 INTEGRATION OVER SUBMANIFOLDS OF RN 176
Thus, for each Ω ⊆ Rn be open and bounded, having a C 1 -boundary, Th. 3.30
yields Z Z
n n
n β (Ω) = div F dλ = hx, ν(x)i dσ (x). (3.53)
Ω ∂Ω
In particular, for the Euclidean ball Br (0), r > 0, we obtain (writing ωn = β n (B1 (0))
as before)
Z Z
n n (3.43) kxk22
n ωn r = n β (Br (0)) = hx, ν(x)i dσ (x) = dσ (x) = r σ(Sr (0)),
Sr (0) Sr (0) r
recovering (3.31) and (3.32), i.e. σ(S1 (0)) = n ωn and σ(Sr (0)) = σ(S1 (0)) rn−1 .
R
(b) To evaluate the integral S1 (0) (x4 + y 4 ) dσ (x, y), we apply (3.52) with
F : R2 −→ R2 , F (x, y) = (x3 , y 3 ),
div F : R2 −→ R, div F (x, y) = 3x2 + 3y 2 .
Since S1 (0) = ∂B1 (0), we can use (3.43) to obtain ν(x, y) = (x, y) for (x, y) ∈ S1 (0).
Thus,
Z Z D E
4 4
(x + y ) dσ (x, y) = (x3 , y 3 ), (x, y) dσ (x, y)
S1 (0) S (0)
Z 1
(3.52)
= (3x2 + 3y 2 ) dλ2 (x, y)
B1 (0)
Z 1 Z 2π
3 · 2π 3π
= 3 r3 dϕ dr = = ,
0 0 4 2
where we used polar coordinates to evaluate the integral over B1 (0).
—
Finally, as another application of Th. 3.30, we will prove Green’s identities, which, for
example, are of use in the theory of partial differential equations.
Theorem 3.32 (Green’s Identities). Let n ∈ N, n ≥ 2. Let Ω ⊆ Rn be open and
bounded, having a C 1 -boundary (then Ω and ∂Ω are compact; in particular, ∂Ω is a
finitely coverable (n − 1)-dimensional submanifold of class C 1 ). If Ω ⊆ U ⊆ Rn with U
open and f, g ∈ C 2 (U, R), then the following identities hold:
Z Z Z
n
h∇ f, ∇ gi dλ = f h∇ g, νi dσ − f ∆g dλn , (3.54a)
Ω ∂Ω Ω
Z Z
n
(f ∆g − g ∆f ) dλ = f h∇ g, νi − g h∇ f, νi dσ , (3.54b)
Ω ∂Ω
Pn
recalling ∆f = i=1 ∂i ∂i f .
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
A ORDER ON, ARITHMETIC IN, AND TOPOLOGY OF R 177
One has
Indeed,
n
X n
X
h∇ f, ∇ gi + f ∆g = ∂i f ∂i g + f ∂i ∂i g = ∂i f ∂i g = div F (x).
i=1 i=1
Thus,
Z Z Z Z
n n (3.52)
h∇ f, ∇ gi + f ∆g dλ = div F dλ = hF, νi dσ = f h∇ g, νi dσ ,
Ω Ω ∂Ω ∂Ω
proving (3.54a).
If one interchanges the roles of f and g in (3.54a) and subtracts the resulting equation
from the original one, then one obtains (3.54b).
One defines the order on, arithmetic in, and topology on R such that they, as far as
possible, constitute extensions of the respective (standard) definitions on R. In [Phi16b,
Ex. 1.52(b)], we had actually already considered the order and (resulting order) topology
on R. Here, we include it again for convenience.
Definition A.2. (a) The (total) order on R is extended to R by setting −∞ < a < ∞
for each a ∈ R. The absolute value function is extended from R to R by defining
|∞| := | − ∞| := ∞.
(b) The set of open intervals I o in R is defined to consist of R plus all open intervals
in R plus all intervals of the form [−∞, a[ or ]a, ∞], a ∈ R:
I o := R ∪ ]a, b[: a, b ∈ R, a < b
∪ [−∞, a[: a ∈ R ∪ {∞} ∪ ]a, ∞] : a ∈ R ∪ {−∞} . (A.1)
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
A ORDER ON, ARITHMETIC IN, AND TOPOLOGY OF R 178
Then the (standard) topology on R is defined by calling a set O ⊆ R open if, and
only if, each element x ∈ O is contained in an open interval I ∈ I o that is contained
in O, i.e. x ∈ I ⊆ O. In other words, I o is defined to be a base for the topology on
R.
However, the following Lem. A.4 shows that commutativity of addition and multiplica-
tion hold on R as well as associativity of multiplication; associativity of addition also
holds if one removes one of the infinities. Moreover, (R, +) is not a group, as + is not
associative; (R \ {0}, ·) is not a group, as the infinities do not have multiplicative in-
verses; (R \ {∞}, +) and (R \ {−∞}, +) are not groups, as, in both cases, the remaining
infinity does not have an additive inverse.
Lemma A.4. Consider R with addition and multiplication as defined in Def. A.2(c).
a + b = b + a. (A.6)
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
B ALGEBRAIC STRUCTURE ON RINGS OF SUBSETS 179
Proof. (a): If a, b, c ∈] − ∞, ∞], then both sides of (A.5) equal ∞ if, and only if,
∞ ∈ {a, b, c}; if a, b, c ∈ [−∞, ∞[, then both sides of (A.5) equal −∞ if, and only if,
−∞ ∈ {a, b, c}, proving (a).
(b): Commutativity of addition is immediate from (A.2a) and (A.2f).
(c): If 0 ∈ {a, b, c}, then (A.7) holds, since both sides are 0. If a, b, c ∈ R, then (A.7)
holds, since · is associative on R. If 0 ∈ / {a, b, c}, but {−∞, ∞} ∩ {a, b, c} 6= ∅, then
(ab)c = ±∞ and a(bc) = ±∞. However, letting
sgn(±∞) := ±1, (A.9)
due to (A.2d), we obtain
(ab)c = (sgn a · sgn b) · sgn c · ∞ = sgn a · (sgn b · sgn c · ∞ = a(bc),
proving (c).
(d): Commutativity of multiplication is immediate from (A.2d) and (A.2e).
Remark A.5. R with the topology defined in Def. A.2(b) constitutes a so-called com-
pactification of R, i.e. it is a compact topological space such that the topology on R is
recovered as the relative topology when considering R as a subset of R. Actually, it is
straightforward to verify R is homeomorphic to [−1, 1], where
−1
for x = −∞,
x
φ : R −→ [−1, 1], φ(x) := |x|+1 for x ∈ R, (A.10)
1 for x = ∞,
provides a homeomorphism.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
B ALGEBRAIC STRUCTURE ON RINGS OF SUBSETS 180
is bijective. Also recall from Linear Algebra that, with addition and multiplication
modulo 2, Z2 = {0, 1} constitutes a field.
Proof. First note that + and · are associative, commutative, and distributive on {0, 1}X ,
as they are associative, commutative, and distributive on {0, 1}. The neutral element
of addition is χ∅ ≡ 0; the neutral element of multiplication is χX ≡ 1. If f ∈ {0, 1}X ,
then f is its own additive inverse:
(
0 + 0 = 0 if f (x) = 0,
∀ (f + f )(x) =
x∈X 1 + 1 = 0 if f (x) = 1,
completing the proof that ({0, 1}X , +, ·) constitutes a commutative ring with unity.
(a) χA + χB = χA ∆ B .
(b) χA · χB = χA∩B .
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
B ALGEBRAIC STRUCTURE ON RINGS OF SUBSETS 181
Proof. (a) is immediate from (B.1), since ∪ is commutative. Rules (b) – (d) can be
proved concisely by making use of Prop. B.2 and Prop. B.3: If we let X := A ∪ B ∪ C
then A, B, C ∈ P(X) and we are in the situation of the propositions. Let x ∈ X. For
(b), we compute
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
C INITIAL AND FINAL σ-ALGEBRAS 182
Proposition B.5. Let X be a set, E ⊆ P(X). Then (E, ∆, ∩) is a ring in the sense
of Def. 1.13(b) if, and only if, it is a ring in the sense of algebra (i.e. in the sense of
[Phi16a, Def. C.7]).
Remark B.6. The reason that one can not replace ∆ by ∪ in Prop. B.5 lies in the fact
of missing inverse elements: While ∅ is still neutral for ∪, if A 6= ∅, then there does not
exist a set B such that A ∪ B = ∅.
Definition C.1. Let X be a set and let (Xi , Ai )i∈I be a family of measurable spaces,
I 6= ∅.
(a) Given a family of functions (fi )i∈I , fi : X −→ Xi , the initial σ-algebra on X with
respect to the family (fi )i∈I is the smallest σ-algebra A on X that makes all fi
A-Ai -measurable (i.e. A is the intersection of all σ-algebras on X that make all
fi measurable – this intersection is well-defined, since the σ-algebra P(X) on X
always makes all fi measurable).
(b) Given a family of functions (fi )i∈I , fi : Xi −→ X, the final σ-algebra on X with
respect to the family (fi )i∈I is
−1
A := A ⊆ X : ∀ fi (A) ∈ Ai . (C.1)
i∈I
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
C INITIAL AND FINAL σ-ALGEBRAS 183
We will see in Lem. C.2(b) below (in generalization of Prop. 1.55(a)) that A is,
indeed, a σ-algebra, and that it is the largest σ-algebra on X that makes all fi
measurable.1
Lemma C.2. Let X be a set and let (Xi , Ai )i∈I be a family of measurable spaces, I 6= ∅.
is a generator of the initial σ-algebra A on X with respect to the family (fi )i∈I .
(b) Given a family of functions (fi )i∈I , fi : Xi −→ X, the final σ-algebra A on X with
respect to the family (fi )i∈I as defined in (C.1) is, indeed, a σ-algebra on X, and it
is the largest σ-algebra on X that makes all fi measurable.
Proof. (a) is immediate, since each σ-algebra A on X such that all fi , i ∈ I, are A-Ai -
measurable must contain E.
(b): For each i ∈ I, since ∅ = fi−1 (∅) ∈ Ai , one has ∅ ∈ A. If A ∈ A, then X \ A ∈ A,
since, for each i ∈ I, fi−1 (X \ A) = Xi \ fi−1 (A) ∈ Ai , due to Ai being a σ-algebra. If
(An )n∈N is a sequence of sets in A, then, for each i ∈ I,
∞
! ∞
[ [
−1
fi An = fi−1 (An ) ∈ Ai ,
n=1 n=1
S∞
as Ai is a σ-algebra. Thus, n=1 An ∈ A, which completes the proof that A is a
σ-algebra on X. It is immediate from the definition of A that each fi , i ∈ I, is Ai -A-
measurable. To see that A is the largest σ-algebra on X with this property, we still
need to show that every σ-algebra B on X making all fi Ai -B-measurable is contained
in A. To this end, let B be such a σ-algebra on X. If B ∈ B, then, for each i ∈ I,
fi−1 (B) ∈ Ai , i.e. B ∈ A, showing B ⊆ A.
Proposition C.3. Let X be a set and let (Xi , Ai )i∈I be a family of measurable spaces,
I 6= ∅.
(a) Given a family of functions (fi )i∈I , fi : X −→ Xi , let A denote the initial σ-
algebra on X with respect to the family (fi )i∈I . Then A has the property that each
map g : Z −→ X from a measurable space (Z, AZ ) into X is measurable if, and
only if, each map (fi ◦ g) : Z −→ Xi is measurable. Moreover, A is the only
σ-algebra on X with this property.
1
In the language of so-called Category Theory, we can say that the category of measurable spaces
(analogous to the category of topological spaces) has initial and final objects.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
C INITIAL AND FINAL σ-ALGEBRAS 184
(b) Given a family of functions (fi )i∈I , fi : Xi −→ X, let A denote the final σ-algebra
on X with respect to the family (fi )i∈I . Then A has the property that each map
g : X −→ Z from X into a measurable space (Z, AZ ) is measurable if, and only if,
each map (g ◦ fi ) : Xi −→ Z is measurable. Moreover, A is the only σ-algebra on
X with this property.
(b) The trace σ-algebra A|B on B ⊆ X, where (X, A) is a measurable space (cf.
Prop. 1.55(c)), is the initial σ-algebra with respect to the identity inclusion map
ι : B −→ X, ι(x) := x: This is also clear from Lem. C.2(a), since
A|B = A ∩ B : A ∈ A = ι−1 (A) : A ∈ A .
(c) In [Phi16b, Ex. F.4(c)], we considered the quotient topology as an example for a
final topology. Analogously, we consider the quotient σ-algebra as an example for
a final σ-algebra (however, in contrast to quotient topologies, quotient σ-algebras
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
D RIEMANN INTEGRAL ON INTERVALS IN RN 185
do not seem to play a particularly important role in the literature): Let (X, A)
be a measurable space and let ∼ be an equivalence relation on X. Moreover, let
Y := X/ ∼= {[x] : x ∈ X} be the corresponding quotient set (i.e. the set of
corresponding equivalence classes). Then the quotient σ-algebra on Y with respect
to ∼, denoted A/ ∼, is defined as the final σ-algebra with respect to the canonical
projection
π : X −→ Y, π(x) := [x].
Thus, by (C.1),
A/ ∼= A ⊆ Y : π −1 (A) ∈ A .
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
D RIEMANN INTEGRAL ON INTERVALS IN RN 186
Definition D.2. Recall the notion of partition for a 1-dimensional interval [a, b] ⊆ R
from [Phi16a, Def. 10.3]. We now generalize this notion to n-dimensional intervals,
n ∈ N: Given an interval I := [a, b] ⊆ Rn , a, b ∈ Rn , a < b, i.e. I = [a, b] = [a1 , b1 ]×· · ·×
[an , bn ], a partition ∆ of I is given by (1-dimensional) partitions ∆k = (xk,0 , . . . , xk,Nk )
of [ak , bk ],Qk ∈ {1, . . . , n}, Nk ∈ N. Given such a partition ∆ of I, for each (k1 , . . . , kn ) ∈
P (∆) := nk=1 {1, . . . , Nk }, define
n
Y
I(j1 ,...,jn ) := [xk,jk −1 , xk,jk ] = [x1,j1 −1 , x1,j1 ] × · · · × [xn,jn −1 , xn,jn ]. (D.4)
k=1
The number
|∆| := max |∆k | : k ∈ {1, . . . , n}
= max xk,l − xk,l−1 : k ∈ {1, . . . , n}, l ∈ {1, . . . , Nk } , (D.5)
t(j1 ,...,jn ) := (t1,j1 , . . . , tn,jn ) ∈ I(j1 ,...,jn ) for each (j1 , . . . , jn ) ∈ P (∆). (D.6)
and
|Ip ∩ Iq | = 0 for each p, q ∈ P (∆) such that p 6= q, (D.8)
since, for p 6= q, Ip ∩ Iq is either empty or it is an interval such that one side consists of
precisely one point. Moreover, as a consequence of (D.7) and (D.8):
X
|I| = |Ip |. (D.9)
p∈P (∆)
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
D RIEMANN INTEGRAL ON INTERVALS IN RN 187
where r(∆, f ) is called the lower Riemann sum and R(∆, f ) is called the upper Riemann
sum associated with ∆ and f . If ∆ is tagged by τ := (tp )p∈P (∆) , then we also define the
intermediate Riemann sum
X
ρ(∆, f ) := f (tp )|Ip |. (D.11c)
p∈P (∆)
(a) Define
J∗ (f, I) := sup r(∆, f ) : ∆ is a partition of I , (D.12a)
J ∗ (f, I) := inf R(∆, f ) : ∆ is a partition of I . (D.12b)
We call J∗ (f, I) the lower Riemann integral of f over I and J ∗ (f, I) the upper
Riemann integral of f over I.
(b) The function f is called Riemann integrable over I if, and only if, J∗ (f, I) = J ∗ (f, I).
If f is Riemann integrable over I, then
Z Z
f (x) dx := f := J∗ (f, I) = J ∗ (f, I) (D.13)
I I
is called the Riemann integral of f over I. The set of all functions f : I −→ R that
are Riemann integrable over I is denoted by R(I, R) or just by R(I).
(c) The function g : I −→ C is called Riemann integrable over I if, and only if, both
Re g and Im g are Riemann integrable. The set of all Riemann integrable functions
g : I −→ C is denoted by R(I, C). If g ∈ R(I, C), then
Z Z Z Z Z
g := Re g, Im g = Re g + i Im g ∈ C (D.14)
I I I I I
is called the Riemann integral of g over I.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
D RIEMANN INTEGRAL ON INTERVALS IN RN 188
(D.11) implies
(b) An example of a function that is not Riemann integrable is given by the n-dimen-
sional version of the Dirichlet function of [Phi16a, Ex. 10.7(b)], i.e. by
(
0 for x ∈ I \ Qn ,
f : I −→ R, f (x) := (D.18)
1 for x ∈ I ∩ Qn .
P
Since r(∆, f ) = 0 and R(∆, f ) = p∈P (∆) |Ip | = |I| for every partition ∆ of I, one
obtains J∗ (f, I) = 0 6= |I| = J ∗ (f, I), showing that f ∈
/ R(I).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
D RIEMANN INTEGRAL ON INTERVALS IN RN 189
i.e. α is the number of interior nodes that occur in the ∆′k . Then, for each partition ∆
of I, the following holds:
where
|I|
φ(I) := max : k ∈ {1, . . . , n} (D.21)
b k − ak
is the maximal volume of the (n − 1)-dimensional faces of I.
Proof. We carry out the proof of (D.20a) – the proof of (D.20b) can be conducted
completely analogous. Consider the case α = 1. Then there is a unique k0 ∈ {1, . . . , n}
such that there exists ξ ∈ ν(∆′k0 )\{ak0 , bk0 }. If ξ ∈ ν(∆k0 ), then ∆+∆′ = ∆, and (D.20a)
is trivially true. If ξ ∈
/ ν(∆k0 ), then xk0 ,l−1 < ξ < xk0 ,l for a suitable l ∈ {1, . . . , Nk0 }.
Recalling the notation from Def. D.2, we let
Pl (∆) := (j1 , . . . , jn ) ∈ P (∆) : jk0 = l , (D.22)
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
D RIEMANN INTEGRAL ON INTERVALS IN RN 190
and
m′p := inf{f (x) : x ∈ Ip′ }, m′′p := inf{f (x) : x ∈ Ip′′ } for each p ∈ Pl (∆). (D.24)
Then we obtain
X
r(∆ + ∆′ , f ) − r(∆, f ) = m′p |Ip′ | + m′′p |Ip′′ | − mp |Ip |
p∈Pl (∆)
X
= (m′p − mp ) |Ip′ | + (m′′p − mp ) |Ip′′ | . (D.25)
p∈Pl (∆)
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
D RIEMANN INTEGRAL ON INTERVALS IN RN 191
for each sequence ofR tagged partitions (∆k )k∈N of I such that limk→∞ |∆k | = 0, then
f ∈ R(I) and α = I f .
First note β ≤ J∗ (f, I) due to the definition of J∗ (f, I). Moreover, (D.34) implies
liml→∞ r(∆kl + ∆′ , f ) = β. Since r(∆′ , f ) ≤ r(∆kl + ∆′ , f ) and ∆′ is arbitrary, we
obtain J∗ (f, I) ≤ β, i.e. J∗ (f, I) = β. Thus, we have shown that every subsequence of
(r(∆k , f ))k∈N converges to β, showing
as claimed. In the same manner, one conducts the proof of J ∗ (f, I) = limk→∞ R(∆k , f ).
Then (D.31a) is immediate from the definition of Riemann integrability, and (D.31b)
follows from (D.31a), since (D.11) implies
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
D RIEMANN INTEGRAL ON INTERVALS IN RN 192
In consequence, for each partition ∆ of I and each ǫ > 0, there are tags τ∗ := (tp,∗ )p∈P (∆)
and τ ∗ := (t∗p )p∈P (∆) such that
Now let (∆k )k∈N be a sequence of partitions of I such that limk→∞ |∆k | = 0. According
to the above, for each ∆k , there are tags τ∗k := (tkp,∗ )p∈P (∆k ) and τ k,∗ := (tk,∗
p )p∈P (∆k ) such
that
k k k 1 k k,∗ k 1
∀ ρ(∆ , τ∗ , f ) < r(∆ , f ) + ∧ ρ(∆ , τ , f ) > R(∆ , f ) − . (D.40)
k∈N k k
Thus,
(D.30) (∗)
J∗ (f, I) = lim r(∆k , f ) = lim ρ(∆k , τ∗k , f ) = α = lim ρ(∆k , τ k,∗ , f )
k→∞ k→∞ k→∞
(∗∗) (D.30)
= lim R(∆k , f ) = J ∗ (f, I), (D.41)
k→∞
where, at (∗) and (∗∗), we used (D.37), (D.40), and the RSandwich theorem [Phi16a, Th.
7.16]. Since (D.41) establishes both f ∈ R(I) and α = I f , the proof is complete.
Proof. Suppose, for each ǫ > 0, there exists a partition ∆ of I such that (D.42) is
satisfied. Then
J ∗ (f, I) − J∗ (f, I) ≤ R(∆, f ) − r(∆, f ) < ǫ, (D.43)
showing J ∗ (f, I) ≤ J∗ (f, I). As the opposite inequality always holds, we have J ∗ (f, I) =
J∗ (f, I), i.e. f ∈ R(I) as claimed. Conversely, if f ∈ R(I) and (∆k )k∈N is a sequence of
partitions of I with limk→∞ |∆k | = 0, then (D.31a) implies that, for each ǫ > 0, there is
N ∈ N such that R(∆k , f ) − r(∆k , f ) < ǫ for each k > N .
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
E WALLIS PRODUCT 193
E Wallis Product
Proposition E.1. For each n ∈ N0 , consider the integral
Z π
I(n) := (sin x)n dx . (E.1)
0
(b) The integrals of (E.1) have the following product representations for each n ∈ N0 :
Z π n
Y
2n 2k − 1
I(2n) = (sin x) dx = π , (E.3a)
0 k=1
2k
Z π Yn
2k
I(2n + 1) = (sin x)2n+1 dx = 2 . (E.3b)
0 k=1
2k + 1
(c) One has the following representation of π as an infinite product, known as the Wallis
product:
∞ n
π Y 2k 2k Y 2k 2k 2 2 4 4 6 6
= · = lim · = · · · · · ··· .
2 k=1 2k − 1 2k + 1 n→∞
k=1
2k − 1 2k + 1 1 3 3 5 5 7
(E.4)
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
F IMPROPER RIEMANN INTEGRAL OF THE SINC FUNCTION 194
and
∀ I(n) ≥ I(n + 1) ≥ 0. (E.11)
n∈N0
Thus,
I(2n) I(2n − 1) 2n + 1
∀ 1≤ ≤ = , (E.12)
n∈N I(2n + 1) I(2n + 1) 2n
implying
n
Y
2n + 1 I(2n) (E.3) π 2k − 1 2k + 1
1 = lim = lim = lim · , (E.13)
n→∞ 2n n→∞ I(2n + 1) 2 n→∞ k=1 2k 2k
proving (c).
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
F IMPROPER RIEMANN INTEGRAL OF THE SINC FUNCTION 195
The sinc function is, clearly, continuous. We have seen in Ex. 2.58(a) that sinc is not
Lebesgue integrable. The goal of this section is to show that sinc is still improperly
Riemann integrable with Z ∞
sin x π
dx = . (F.2)
0 x 2
To this end, consider the function
Z ∞
sin x
+
g : R −→ R, g(t) := e−tx dx . (F.3)
0 x
For each t ∈ R+ , the integrand and its derivative with respect to t are Lebesgue inte-
grable, since they are dominated by the Lebesgue integrable function
(
| sin x|
x
for x < 1,
x 7→ −tx
(F.4)
xe for x ≥ 1.
For each ǫ > 0 and each t ≥ ǫ the integrand in (F.3) is uniformly dominated by
(
| sin x|
x
for x < 1,
x 7→ −ǫx
(F.5)
xe for x ≥ 1.
This shows that Cor. 2.24 applies on [ǫ, ∞[, implying g to be differentiable (even on R+ ,
since ǫ > 0 was arbitrary),
Z ∞ Z ∞
′ ′ −tx sin x
+
g : R −→ R, g (t) := − xe dx = − e−tx sin x dx . (F.6)
0 x 0
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
F IMPROPER RIEMANN INTEGRAL OF THE SINC FUNCTION 196
and Z ∞
∞
sin x X ∞ X
dx = a0,k , ∀ a0,k < |a0,j+1 |. (F.16)
0 x j∈N
k=1 k=j+1
and Z Z Z ∞
∞
sin x r
sin x sin x ǫ
∀ dx −
dx = dx < . (F.18)
r≥R x x x 4
0 0 r
Now let r ≥ R. Using (F.13) and (F.14), we choose δ > 0 such that
Z ∞
π sin x π ǫ
∀ > e−tx dx = g(t) > − (F.19)
0<t<δ 2 0 x 2 4
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
G TOPOLOGICAL AND MEASURE-THEORETIC SUPPLEMENTS 197
and Z Z
r
sin x r
sin x ǫ
∀ dx − e −tx
dx < . (F.20)
0<t<δ x x 4
0 0
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
G TOPOLOGICAL AND MEASURE-THEORETIC SUPPLEMENTS 198
Ṡn
i ∈ {1, . . . , n}, and R \ S = i=1 Ti . Then
[n [
−1 −1 −1 −1 ˙ ˙ n
f (R) \ f (S) = f (R \ S) = f Ti = f −1 (Ti ),
i=1 i=1
Proof. First, assume f ≥ 0 (i.e. f ∈ M+ (B)) and choose a sequence (φk )k∈N in S + (B)
such that φk ↑ f . Then
N
!
X
f = lim φN = lim φ1 + (φk − φk−1 ) . (G.1)
N →∞ N →∞
k=2
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
H POLAR COORDINATES 199
and Z
lim |fn − f | dµ = 0. (G.4b)
n→∞ X
Proof. If µ(X) = 0, there is nothing to prove. Otherwise, given ǫ > 0, there exists
N ∈ N such that, for each n > N , kfn − f k∞ < ǫ/µ(X). Thus, for each n > N ,
Z
ǫ
|fn − f | dµ ≤ · µ(X) = ǫ,
X µ(X)
proving (G.4b). As
Z Z Z
∀ fn dµ − f dµ ≤ |fn − f | dµ ,
n∈N
X X X
H Polar Coordinates
Let
O := R+ ×]0, 2π[, U := R2 \ {(x, 0) : x ≥ 0}. (H.1)
We verify that the function defining polar coordinates, i.e.
constitutes a C 1 -diffeomorphism:
We first show φ to be bijective with inverse map
φ−1 : U −→ O, φ−1 (x, y) = φ−1 −1
1 (x, y), φ2 (x, y) , (H.3a)
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
H POLAR COORDINATES 200
where, recalling the definition of arccot from [Phi16a, Def. and Rem. 8.27],
p
φ−1
1 (x, y) := x2 + y 2 , (H.3b)
arccot(x/y) for y > 0,
−1
φ2 (x, y) := π for y = 0, (H.3c)
π + arccot(x/y) for y < 0.
One verifies φ−1 ◦ φ = IdO :
∀ (φ−1 ◦ φ)(r, ϕ) = φ−1 (r cos ϕ, r sin ϕ) = (r, ϕ),
(r,ϕ)∈O
since
q
φ−1
1 (r cos ϕ, r sin ϕ) = r2 (cos2 ϕ + sin2 ϕ) = r,
arccot(cot ϕ) = ϕ for 0 < ϕ < π,
−1
φ2 (r cos ϕ, r sin ϕ) = π = ϕ for ϕ = π,
π + arccot(cot ϕ) = π + ϕ − π = ϕ for π < ϕ < 2π;
and φ ◦ φ−1 = IdU :
p p
−1
(φ ◦ φ )(x, y) = x2 + y2 cos φ−1
2 (x, y),
2 2 −1
x + y sin φ2 (x, y)
∀
(x,y)∈U
= (x, y),
since
√ 1
= q 1 y2 = √ x2 2 for x > 0 (⇒ y 6= 0),
1+(cot arccot(x/y)) −2
1+ 2 x +y
x
cos(π/2) = 0 = √ 2 2 x
for x = 0, y > 0,
x +y
cos φ−1
2 (x, y) = cos(3π/2) = 0 = √ x2 2 for x = 0, y < 0,
x +y
−1
√ = √ x2 2 for x < 0, y 6= 0,
1+(cot arccot(x/y))−2 x +y
cos π = −1 = √ x2 2 for y = 0 (⇒ x < 0),
x +y
√ 1
= r 1 2 = √ 2y 2 for y > 0,
1+cot 2 arccot(x/y)
1+ x2 x +y
y
sin φ−1
2 (x, y) = sin π = 0 = √ 2y 2 for y = 0,
x +y
√
−1
=√ y for y < 0.
1+cot2 arccot(x/y) x2 +y 2
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06
REFERENCES 201
References
[Alt06] Hans Wilhelm Alt. Lineare Funktionalanalysis, 5th ed. Springer-Verlag,
Berlin, 2006 (German).
[Beh87] E. Behrends. Maß- und Integrationstheorie. Springer-Verlag, Berlin, 1987
(German).
[Els07] Jürgen Elstrodt. Maß- und Integrationstheorie, 5th ed. Grundwissen
Mathematik, Springer-Verlag, Berlin, 2007 (German).
[For17] Otto Forster. Analysis 3, 8th ed. Springer Spektrum, Wiesbaden, Ger-
many, 2017 (German).
[Kön64] H. König. Ein einfacher Beweis des Integralsatzes von Gauß. Jahrestagung
der Deutschen Mathematischen Vereinigung. Vol. 66, 1964 (German).
[Kön04] Konrad Königsberger. Analysis 2, 5th ed. Springer-Verlag, Berlin, 2004
(German).
[Oss09] E. Ossa. Topologie. Vieweg+Teubner, Wiesbaden, Germany, 2009 (German).
[Phi16a] P. Philip. Analysis I: Calculus of One Real Variable. Lecture Notes, LMU Mu-
nich, 2015/2016, AMS Open Math Notes Ref. # OMN:202109.111306, avail-
able in PDF format at
https://www.ams.org/open-math-notes/omn-view-listing?listingId=111306.
[Phi16b] P. Philip. Analysis II: Topology and Differential Calculus of Several Vari-
ables. Lecture Notes, LMU Munich, 2016, AMS Open Math Notes Ref. #
OMN:202109.111307, available in PDF format at
https://www.ams.org/open-math-notes/omn-view-listing?listingId=111307.
[RF10] Halsey Royden and Patrick Fitzpatrick. Real Analysis, 4th ed. Pearson
Education, Boston, USA, 2010.
[Rud87] W. Rudin. Real and Complex Analysis, 3rd ed. McGraw-Hill Book Company,
New York, 1987.
AMS Open Math Notes: Works in Progress; Reference # OMN:202109.111308; Last Revised: 2021-09-17 11:22:06