Chapter 3.

Lebesgue integral and the monotone convergence

theorem
3.1 Starting point
Remark 3.1.1. What we need to dene integrals of R-valued functions, apart from a considerable
amount of the terminology we have come across so far, is that
we have R and the Borel -algebra on it; this will come in for the values of the functions
we consider;
on the domain we need a measure, and a measure needs a -algebra. We will use (R, L, ),
where L is the -algebra of Lebesgue measurable sets and : L [0, ] is the measure
given by (F) = m

(F) for F L.
In fact, we could equally well have a more general domain X and we would need a -algebra
of subsets of X together with a measure : [0, ]. That is we could treat
_
X
f d =
_
X
f(x) d(x), integrals of (certain) functions f : X R with general domains X, provided we
have (X, , ), but we will mostly stick to f : R R where (X, , ) = (R, L, = m

[
L
).
One advantage of sticking to domain X = R is that we can picture our functions as graphs
y = f(x) in R
2
. We could still have that advantage if we allowed X R, for instance to dene
_
[a,b]
f d where f : [a, b] R. Actually this case will be included rather easily, but another
perspective is to generalise to multiple integrals, that is to integrals of functions f : R
n
R. We
are not ready for that because we only did m

as length outer measure on R. For R

2
, we would
need to do something similar to what we did with length to dene area measure. The starting
point would be areas of rectangles, and we would need to replace what we did with the interval
algebra I by stuff about a rectangle algebra. Some of the details at the start would be a bit
different but many of the same proofs can be used in R
2
, and for volume outer measure in R
3
and more generally for n-dimensional measure in R
n
.
However, what we do in this chapter can be done in considerable generality and we set out
the general denitions concisely.
Denition 3.1.2. If X is a set then a collection T(X) is called a -algebra (of subsets of
X) if it satises:
(i)
(ii) E E
c

(iii) E
1
, E
2
, . . .

n=1
E
n
.
A pair (X, ) of a set X and a -algebra of subsets of X is sometimes called a measurable space.
Denition 3.1.3. If is a -algebra (of subsets of set X) and : [0, ] is a function, then
we call a measure on if it satises
2 201112 Mathematics MA2224
(a) () = 0
(b) is countably additive, that is whenever E
1
, E
2
, . . . are disjoint, then (

n=1
E
n
) =

n=1
(E
n
).
The combination (X, , ) is called a measure space and the subsets E X with E are
called measurable subsets (with respect to the given ). If in addition (X) = 1, then is called
a probability measure and the combination (X, , ) is called a probability space. In the context
of probability, measurable subsets E are called events and (E) [0, 1] the probability of
the event E.
Examples 3.1.4. 1. Our primary focus will be (X, , ) = (R, L, ) with = m

[
L
being
Lebesgue length measure on the -algebra L of Lebesgue measurable subsets of R.
2. If X L then we can dene a -algebra on X and a measure : R by
= F X : F L
and
(F X) = (F X) = m

(F X).
It may be helpful to have a notation L
X
for this (though that is not a standard notation).
Proof. There is nothing difcult in this except perhaps that there is a possibility of confu-
sion between two notions of complement. Perhaps this can be solved if we write R E for
the complement of E R and XE for the complement of E X as a subset of X. The
notation we usually use of E
c
for the complement of E is not very adaptable to more than
one meaning of complement.
Notice that, since X L, we could also say
= E L : E X
(because E L with E X implies E = E X, and on the other hand F L implies
that E = F X L and has E X). So , E X E = X (R E) .
Moreover E
1
, E
2
, . . . implies

n=1
E
n
.
The properties required for to be a measure are just true because we know is a measure.
3. If X = [0, 1], then the previous example turns into an example of a probability space
([0, 1], , ).
4. If (X, ,
1
) and (X, ,
2
) are two measure spaces with the same underlying set and
-algebra, and if c
1
, c
2
0, then we can get a new measure = c
1

1
+ c
2

2
on by
dening
(E) = c
1

1
(E) +c
2

2
(E) (E )
(This is easy to check and we can leave it as an exercise.)
Lebesgue integral 3
3.2 Measurable functions
Denition 3.2.1. A function f : R R is called Lebesgue measurable (or measurable with
respect to the measurable space (R, L)) if for each a R
x R : f(x) a = f
1
((, a]) L.
Examples 3.2.2. (i) Continuous functions f : R Rare always measurable, because (, a]
is a closed subset and so f
1
((, a]) is closed, hence in L.
If you prefer to think about open sets, (a, ) is open, so f
1
((a, )) is open, hence in L,
hence its (closed) complement (f
1
((a, )))
c
= f
1
((a, )
c
) = f
1
((, a]) is in L.
(ii) The characteristic function
Q
of the rationals is measurable (but not continuous anywhere).
The reason is that the set Q L (it is a countable union of single point sets, hence a
Borel set, hence in L; another way is to use m

(Q) = 0 to get Q L) and so then is its

complement R Q L. For f =
Q
, we can see that
f
1
((, a]) = x R : f(x) a =
_

_
if a < 0
R Q if 0 a < 1
R if a 1
and in all cases we get a set in L.
(iii) For a subset E R, its characteristic function
E
is a measurable function if and only if
E L (which means E is a measurable set). As in the previous case, for f =
E
, we can
see that
f
1
((, a]) = x R : f(x) a =
_

_
if a < 0
R E if 0 a < 1
R if a 1
Certainly , R L and we have R E L E L.
Remark 3.2.3. Our aim will be to dene integrals for all measurable functions, but we will not
quite succeed in that.
Proposition 3.2.4. If f : R R is Lebesgue measurable, then
f
1
(B) L
for each Borel set B.
Proof. Let

f
= E R : f
1
(E) L.
We claim that
f
is a -algebra.
That is quite easy to verify.
4 201112 Mathematics MA2224
(i) f
1
() = L
f
(ii) E
f
f
1
(E) L f
1
(E
c
) = (f
1
(E))
c
L E
c

f
(iii) E
1
, E
2
, . . .
f
f
1
(E
n
) L for n = 1, 2, . . .. From this we have

n=1
f
1
(E
n
)
L and so
f
1
_

_
n=1
E
n
_
=

_
n=1
f
1
(E
n
) L

_
n=1
E
n

f
As we know f is Lebesgue measurable, we have
(, a]
f
(a R)
and that means that
f
contains the -algebra generated by (, a] : a R which we
know to be the Borel -algebra by Corollary 2.3.12.
Proposition 3.2.5. If f, g : R R are Lebesgue measurable functions and c R, then the
following are also Lebesgue measurable functions
cf, f
2
, f +g, fg, [f[, max(f, g)
The idea here is to combine functions by manipulating their values at a point. So fg : R R
is the function with value at x R given by (fg)(x) = f(x)g(x), and similarly for the other
functions.
Proof. cf: First if c = 0, cf is the zero function, which is measurable (very easy to check that
directly). Let h = cf (so that h(x) = cf(x)). For c > 0, then
h
1
((, a]) = x R : cf(x) a = x R : f(x) a/c L
and so h = cf is Lebesgue measurable. For c < 0,
h
1
((, a]) = x R : f(x) a/c = f
1
([a/c, )) L
using Proposition 3.2.4 and the fact that [a/c, ) is a Borel set.
f
2
: Here we say
x R : f(x)
2
a =
_
if a < 0
f
1
([

a,

a]) if a 0
f +g: (This is perhaps the most tricky part.) Fix a R and we aim to show that x R :
f(x) + g(x) a L. Taking the complement, this would follow from x R :
f(x) +g(x) > a L.
For any x where f(x) + g(x) > a then f(x) > a g(x) and there is a rational q Q so
that f(x) > q > a g(x). Then g(x) > a q. So
x S
q
= f
1
((q, )) g
1
((a q, ))
Lebesgue integral 5
for some q Q. On the other hand if f(x) > q and g(x) > a q, then f(x) + g(x) > a,
from which we conclude that
x R : f(x) +g(x) > a =
_
qQ
S
q
.
By Proposition 3.2.4 S
q
L for each q and hence x R : f(x) + g(x) > a L
(because it is countable union of sets S
q
L).
fg: This follows from the previous parts because fg = ((f +g)
2
(f g)
2
)/4.
[f[: This is easy as x R : [f(x)[ a = if a < 0 but is f
1
([a, a]) if a 0.
max(f, g): x R : max(f(x), g(x)) a = f
1
((, a]) g
1
((, a]).
Remark 3.2.6. It is convenient to allow for functions that sometimes have the value and
sometimes . For that we need to introduce the extended real line [, ], which is R
with two extra elements and +added. This is similar to our earlier use of [0, ].
We order [, ] so that < x < for each x R. We introduce arithmetic on
[, ] in a similar way to what we did before for [0, ] but we do not allow subtraction of
from itself or from itself. We extend the usual arithmetic operations on R with the rules
x + = +x =
x + () = () +x = () (for x R)
x = x =
x() = ()x = () (for 0 ,= x R)
0 = 0 = 0
0() = ()0 = 0
From the point of view of the ordering of [, ], there is a strictly monotone increasing
bijection : R (1, 1) given by
(x) =
x
1 +[x[
and both and
1
are continuous (so that is a homeomorphism). We can extend to get an
order-preserving bijection : [, ] [1, 1] by dening () = 1, (x) = (x) for
x R and () = 1. Since we know that nonempty subsets of [1, 1] have both a supremum
and an inmum, it follows by applying
1
that all nonempty subsets of [, ] also have a
supremum (least upper bound) and inmum (greatest lower bound).
We dene convergence on [, ] such that becomes a homeomorphism onto [1, 1]
(which means for example that lim
n
x
n
= has the standard meaning if all the terms x
n
and
the limit are nite, but lim
n
x
n
= means lim
n
(x
n
) = () = 1 or that given any
M < , there is N so that M < x
n
holds for all n > N). It is also true that monotone
increasing sequences (x
n
)

n=1
in [, ] always converge (to sup
n
x
n
in fact) and so also do
monotone decreasing sequences always converge (to their inmum).
6 201112 Mathematics MA2224
Denition 3.2.7. A function f : R [, ] is called Lebesgue measurable (or measurable
with respect to the measurable space (R, L)) if for each a R
x R : f(x) a = f
1
([, a]) L.
Note that this denition agrees with the earlier Denition 3.2.1 if f(x) R always (because
then f
1
([, a]) = f
1
((, a])).
Proposition 3.2.8. Let f : R [, ] be a function and let A = x R : f(x) = =
f
1
(), B = x R : f(x) = = f
1
(). Then the following are equivalent
(a) f is Lebesgue measurable
(b) A, B L and the function g : R R given by
g(x) =
_
f(x) if < x <
0 if x A B
is Lebesgue measurable.
Proof. (a) (b): Note that
A =

n=1
f
1
([, n]) L
(because -algebras are closed under the operation of taking countable intersections) and
R B =

_
n=1
f
1
([, n]) L B L
(countable unions and complements stay in L).
To show that g is measurable, note that
g
1
((, a]) =
_
f
1
([, a]) B if a 0
f
1
([, a]) A if a < 0
So g
1
((, a]) L for each a.
(b) (a): Note that
f
1
([, a]) =
_
g
1
((, a]) B if a 0
g
1
((, a]) A if a < 0
and so f
1
([, a]) L for each a (using (b)).
Remark 3.2.9. We can also dene measurability of functions dened on Lebesgue measurable
subsets X R, as follows. This would include for example the situation where X is an interval
of any kind, or any Borel set (which includes open subsets X of R, in particular).
Lebesgue integral 7
Denition 3.2.10. If X L, then a function f : X [, ] is called Lebesgue measurable
on X if for each a R
x X : f(x) a = f
1
([, a]) L.
Proposition 3.2.11. Let X L and f : R [, ]. Then the following are equivalent
(a) f is Lebesgue measurable on X
(b) the function g : R [, ] given by
g(x) =
_
f(x) if x X
0 if x R X
is Lebesgue measurable on R.
Proof. For a R we have
g
1
([, a]) = x R : g(x) a
=
_
(R X) x X : f(x) a if a 0
x X : f(x) a if a < 0
So (a) (b). For the converse use
f
1
([, a]) =
_
x R : g(x) a (R X) if a 0
x R : g(x) a if a < 0
We can extend Proposition 3.2.5 to functions with values in [, ] except for one dif-
culty.
Proposition 3.2.12. If X L, c R and f, g : R [, ] are Lebesgue measurable
functions on X, then the following are also Lebesgue measurable functions (on X)
cf, f
2
, fg, [f[, max(f, g)
If x : f(x) = x : g(x) = = = x : f(x) = x : g(x) = , then f +g
is measurable (on X).
The proof (which we leave as an exercise) is to use Propositions 3.2.11, 3.2.8, and 3.2.5 for
cf (c ,= 0), f
2
, f +g, fg and [f[. For max(f, g) a similar proof can be used.
8 201112 Mathematics MA2224
3.3 Limits of measurable functions
The next result shows that taking (pointwise) limits of sequence of measurable functions still
give a measurable limit.
Denition 3.3.1. If f
n
: R [, ] (n = 1, 2, . . .) is an innite sequence of functions, then
we say that f : R [, ] is the pointwise limit of the sequence (f
n
)
n
if we have
f(x) = lim
n
f
n
(x)
for each x R.
For any sequence f
n
: R [, ] we can dene limsup
n
f
n
as the function with value
at x given by
limsup
n
f
n
(x) = lim
n
_
sup
kn
f
k
(x)
_
(something that always makes sense because sup
kn
f
k
(x) decreases as n increases or at least
does not get any bigger as n increases).
Proposition 3.3.2. Let f
n
: R [, ] (n = 1, 2, . . .) be Lebesgue measurable functions.
Then
(a) the function x sup
n1
f
n
(x) is Lebesgue measurable;
(b) the function x inf
n1
f
n
(x) is Lebesgue measurable;
(c) limsup
n
f
n
is Lebesgue measurable; and
(d) if the pointwise limit function f : R [, ] exists, then f is Lebesgue measurable.
Proof. (a) If g(x) = sup
n1
f
n
(x), then g
1
([, a]) =

n=1
f
1
n
([, a]) L (since the
-algebra L is closed under taking countable intersections).
(b) If h(x) = inf
n1
f
n
(x), then g(x) = sup
n1
(f
n
(x)) is measurable by (a) and so then is
g(x) measurable.
(c) We have limsup
n
f
n
(x) = inf
n1
g
n
(x) where g
n
(x) = sup
kn
f
k
(x) is a decreasing
sequence. But each g
n
is measurable by (b) and then limsup
n
f
n
(x) is measurable by
(a).
(d) If the pointwise limit f(x) exists, then f(x) = limsup
n
f
n
(x), which is measurable by
(c).
Lebesgue integral 9
3.4 Simple functions
Denition 3.4.1. A function f : R R is called a simple function if the range f(R) is a nite
set.
Note that we require nite values.
Proposition 3.4.2. Each simple function f : R R has a representation
f =
n

i=1
a
i

E
i
as a linear combination of nitely many characteristic functions of disjoint sets E
1
, E
2
, . . . , E
n
with coefcients a
1
, a
2
, . . . , a
n
R.
Moreover we can assume a
1
, a
2
, . . . , a
n
are distinct values, that the E
j
are all nonempty and
that

n
j=1
E
j
= R. With these assumptions the representation is called the standard representa-
tion and is unique apart from the order of the sum.
A simple function is Lebesgue measurable if and only if the sets E
j
in the standard represen-
tation are all in L.
Proof. Since f(R) is a nite set (and cant be empty) we can list the elements f(R) = y
1
, y
2
, . . . , y
n

with n 1. Take E
j
= f
1
(y
j
) = x R : f(x) = y
j
and then we get the standard repre-
sentation f =

n
i=1
y
i

E
i
.
If f is measurable then E
j
L for each j (because one point sets a
j
are Borel sets and
we can use Proposition 3.2.4).
Conversely, if each E
j
is in L, then each
E
j
measurable, and then so is f measurable by
Proposition 3.2.5.
Notation 3.4.3. We will develop integration rst for nonnegative measurable functions f : R
[0, ].
To reduce the case of general (Lebesgue measurable) f : R [, ] to the positive case
we will use the two associated functions f
+
: R [0, ] and f

: R [0, ] given by
f
+
= max(f, 0) =
1
2
([f[ +f), f

= max(f, 0) =
1
2
([f[ f),
both of which we know to be measurable. Notice that f = f
+
f

and [f[ = f
+
+f

.
We call f
+
the positive part of f and f

the negative part (but note that f

is positive also).
Denition 3.4.4. If f : R [0, ) is a nonnegative measurable simple function, with standard
representation f =

n
j=1
a
j

E
j
, then we dene
_
R
f d =
n

j=1
a
j
(E
j
)
(where (E
j
) means the Lebesgue measure of E
j
). (Note that the integral makes sense in [0, ].)
10 201112 Mathematics MA2224
For E L, we dene
_
E
f d =
_
R

E
f d
(noting that
E
f is also a nonnegative measurable simple function).
Examples 3.4.5. (i) For f =
[0,1]
+2
[4,)
the possible values are 0, 1 and 2, and the standard
representation is
f =
[0,1]
+ 2
[4,)
+ 0
(,0)(1,4)
so that
_
R
f d = 1([0, 1]) + 2([4, )) + 0((, 0) (1, 4)) = 1 ++ 0 =
(ii) For f =
Q
, the standard representation is f =
Q
+ 0
R\Q
and
_
R
f d = 1(Q) + 0(R Q) = 0 + 0 = 0
(iii) (Exercise) Show that if E L, then
_
R

E
d = (E).
Proposition 3.4.6. If f, g : R [0, ) are measurable simple functions and [0, ), then
f, f +g are measurable simple functions and
_
R
f d =
_
R
f d,
_
R
(f +g) d =
_
R
f d +
_
R
g d.
Proof. We know from Proposition 3.2.5 that f, f +g are measurable.
If = 0, then 0f has standard representation 0 = 0
R
and integral
_
R
f d =
_
R
0
R
d =
0(R) = 0 = 0
_
R
f d =
_
R
f d. If > 0 and f(R) = a
1
, a
2
, . . . , a
n
let F
j
= f
1
(a
j
)
so that f has standard representation f =

n
j=1
a
j

F
j
, while f has standard representation
f =

n
j=1
a
j

F
j
. So
_
R
f d =
n

j=1
(a
j
)(F
j
) =
n

j=1
a
j
(F
j
) =
_
R
f d.
To cope with h = f+g, let f(R) = a
1
, a
2
, . . . , a
n
, F
j
= f
1
(a
j
), g(R) = b
1
, b
2
, . . . , b
m
,
G
k
= g
1
(b
k
). The range of h = f + g is certainly nite because it is contained in a
j
+ b
k
:
1 j n, 1 k m. So h is a simple function. Write h(R) = c
1
, c
2
, . . . , c
p
and then for
each 1 p we have
H

= h
1
(c

) =
_
{(j,k):a
j
+b
k
=c

}
F
j
G
k
.
The sets making up H

are a disjoint union and so note that nite additivity of gives us

(H

) =

{(j,k):a
j
+b
k
=c

}
(F
j
G
k
)
Lebesgue integral 11
We now have standard representations f =

n
j=1
a
j

F
j
, g =

m
k=1
b
k

G
k
, h = f + g =

p
=1
c

and
_
R
hd =
p

=1
c

(H

)
=
p

=1
c

{(j,k):a
j
+b
k
=c

}
(F
j
G
k
)
=

{(j,k):F
j
G
k
=}
(a
j
+b
k
)(F
j
G
k
)
=

1jn,1km
(a
j
+b
k
)(F
j
G
k
)
(since () = 0)
=
n

j=1
m

k=1
a
j
(F
j
G
k
) +
m

k=1
n

j=1
b
k
(F
j
G
k
)
=
n

j=1
a
j

_
m
_
k=1
F
j
G
k
_
+
m

k=1
b
k

_
n
_
j=1
F
j
G
k
_
(using nite additivity of )
=
n

j=1
a
j
(F
j
) +
m

k=1
b
k
(G
k
)
=
_
R
f d +
_
R
g d
Remark 3.4.7. The tricky part in the above proof is to stick to the denition of the integral of a
simple function in terms of the standard representation. We do this so that the denition is not
ambiguous, but now that we have proved this result we can say that if we have any representation
of a simple function f =

n
i=1
y
j

E
j
with y
j
0 and E
j
measurable, then
_
R
f d =
n

i=1
y
j
_
R

E
j
d =
n

i=1
y
j
(E
j
)
Corollary 3.4.8. If f, g : R [0, ) are measurable simple functions with f(x) g(x) for
each x, then
_
R
f d
_
R
g d
Proof. h = g f is measurable (by Proposition 3.2.5), simple (can only have nitely many
values) and also h(x) 0. So
_
R
hd is dened and by Proposition 3.4.6,
_
R
g d =
_
R
(f +h) d =
_
R
f d +
_
R
hd
_
R
f d.
12 201112 Mathematics MA2224
Proposition 3.4.9. If f : R [0, ) is a measurable simple function, then we can dene a
measure
f
: L [0, ] by

f
(E) =
_
E
f d
Proof. We know we can write f in the form f =

n
j=1
a
j

E
j
with a
j
0 and E
j
L
(1 j n). So

E
f =
n

j=1
a
j

E
j
=
n

j=1
a
j

EE
j
and

f
(E) =
_
E
f d =
_
R

E
f d =
n

j=1
a
j
(E E
j
)
We know that
E
j
(E) = (EE
j
) denes a measure on L (Tutorial 6, Q2) and it follows from
Examples 3.1.4-4 that
f
=

n
j=1
a
j

E
j
is a measure.
Proposition 3.4.10. If : L [0, ] is a measure, then it satises
(a) If E
1
, E
2
, . . . L with E
1
E
2
(that is a monotone increasing sequence of
measurable sets) then

_
n=1
E
n
_
= lim
n
(E
n
) = sup
n
(E
n
)
(b) If E
1
, E
2
, . . . L with E
1
E
2
(that is a monotone decreasing sequence of
measurable sets) and if (E
1
) < then

n=1
E
n
_
= lim
n
(E
n
) = inf
n
(E
n
)
Proof. (a) (See similar proof on tutorial sheet 5 (for the special case = = m

. Also look at a
Venn diagram to see what is happening.) If (E
n
)

n=1
is a monotone increasing sequence in L
then we can take F
1
= E
1
, F
2
= E
2
E
1
, F
3
= E
3
E
2
, in general F
n
= E
n
E
n1
for n 2,
so that F
n
/for each n, (F
n
)

n=1
a disjoint sequence (because E
1
E
2
E
3
means
that F
n
= E
n

_
n1
i=1
E
i
_
and so F
n
E
i
= for i < n, and that gives F
n
F
i
= as
F
i
E
i
).
Now
n
_
i=1
F
i
= E
n
for each n and

_
i=1
F
i
=

_
n=1
E
n
.
Lebesgue integral 13
By countable additivity of (and disjointness of the F
i
),

_
i=1
E
i
_
=
_

_
i=1
F
i
_
=

i=1
(F
i
) = sup
n
n

i=1
(F
i
) = sup
n

_
n
_
i=1
F
i
_
= sup
n
(E
n
)
(b) Assume now that (E
n
)

n=1
is a monotone decreasing sequence in L. Let G
n
= E
1
E
n
so
that E
1
= G
n
E
n
, a disjoint union of sets on /, and
(E
1
) = (G
n
) +(E
n
).
Using (E
1
) < , this implies
(E
n
) = (E
1
) (G
n
).
Since G
1
G
2
G
3
we have

_
n=1
G
n
_
= lim
n
(G
n
) = sup
n
(G
n
)
by (a). Now,
E
1

n=1
E
n
= E
1

n=1
E
n
_
c
=

_
n=1
(E
1
E
c
n
) =

_
n=1
(E
1
E
n
) =

_
n=1
G
n
and, as above,
(E
1
) =
_
E
1

n=1
E
n
_
+
_

n=1
E
n
_
=
_

_
n=1
G
n
_
+
_

n=1
E
n
_

n=1
E
n
_
= (E
1
)
_

_
n=1
G
n
_
= (E
1
) lim
n
(G
n
)
= lim
n
(E
1
) (G
n
) = lim
n
(E
n
)
Since (E
1
) (E
2
) , we have lim
n
(E
n
) = inf
n
(E
n
).
3.5 Positive measurable functions
Proposition 3.5.1. If f : R [0, ] is a (nonnegative extended real valued) measurable func-
tion, then there is a monotone increasing sequence (f
n
)

n=1
of (nonnegative) measurable simple
functions f
n
: R [0, ) with
lim
n
f
n
(x) = f(x) (for each x R)
14 201112 Mathematics MA2224
Proof. For each n, dene
E
n,j
= f
1
__
j 1
2
n
,
j
2
n
__
1 j n2
n
and
E
n,1+n2
n = f
1
([n, ]).
Note that each E
n,j
is in L. Then put
f
n
=
n2
n
+1

j=1
j 1
2
n

E
n,j
.
Observe that f
n
(x) f(x) always. For a given x where f(x) < we have f
n
(x) = n for
n f(x) but for n > f(x) we nd that
f
n
(x) =
1
2
n
2
n
f(x)|
where y| means the greatest integer y. Since 2
n+1
f(x)| 22
n
f(x)|, it follows that
f
n+1
(x) f
n
(x) always.
Also as 2
n
f(x)| 2
n
f(x) 1 we get f
n
(x) > f(x)
1
2
n
as soon as n > f(x). Since
f
n
(x) f(x) we get lim
n
f
n
(x) = f(x) if f(x) < .
Finally, if f(x) = , then f
n
(x) = n for all n and so lim
n
f
n
(x) = = f(x) in this
case also.
Remark 3.5.2. It is tempting to dene
_
R
f d (for f : R [0, ] a measurable function) as
lim
n
_
R
f
n
d (where f
n
is exactly as in the proof of the above proposition). But that makes
it a little tricky to prove that the integral has the various obvious properties it has. So we use a
more complicated denition rst and prove (as part of the monotone convergence theorem) that
we get the same value for the integral.
Another thing to watch out for is that we dont change the denition we already have when
f : R R is a measurable simple function.
Denition 3.5.3. For f : R [0, ] a Lebesgue measurable function, we dene
_
R
f d = sup
__
R
s d : s: R [0, ) a measurable simple function with s f
_
For E L, we dene
_
E
f d =
_
R

E
f d
(noting that
E
f is also a nonnegative measurable function).
Remark 3.5.4. By Corollary 3.4.8, when f : R R is itself a simple measurable function, this
gives the same value for
_
R
f d (and
_
E
f d) as in the earlier denition (Denition 3.4.4).
Lebesgue integral 15
Lemma 3.5.5. For f, g : R [0, ] Lebesgue measurable functions, we have
(a) if f g (pointwise) then
_
R
f d
_
R
g d
(b) if E F with E, F L, then
_
E
f d
_
F
f d
Proof. (a) if f g, then any simple function s: R [0, ) with s f also has s g and so
_
R
s d
_
R
g d. Taking the supremum over all such s we get the result.
(b) Since E F, we have
E
f
F
f and so
_
E
f d =
_
R

E
f d
_
R

F
f d =
_
F
f d
by (a).
Theorem 3.5.6 (Monotone Convergence Theorem). If f
n
: R [0, ] is a monotone increasing
sequence of (Lebesgue) measurable functions with pointwise limit f, then
_
R
f d = lim
n
_
R
f
n
d
Proof. From f
n
f
n+1
we have f
j
(x) lim
n
f
n
(x) = f(x) for each j. Hence
_
R
f
j
d
_
R
f d for each j by Lemma 3.5.5. Also
_
R
f
n
d
_
R
f
n+1
d and so the sequence of integrals
_
R
f
n
d is monotone increasing. That means lim
n
_
R
f
n
d makes sense in [0, ]. We have
lim
n
_
R
f
n
d = sup
n1
_
R
f
n
d
_
R
f d.
It remains to prove the reverse inequality.
Fix a simple function s: R Rwith s(x) f(x) for all x R. We will showlim
n
_
R
f
n
d
_
R
s d and that will be enough because of the way
_
R
f d is dened.
To show this, x (0, 1) and consider
E
n,
= x R : f
n
(x) s(x)
Notice that
_
En,
s d
_
En,
f
n
d
_
R
f
n
d (1)
Because f
n+1
f
n
we know E
n,
E
n+1,
. For s(x) = 0, x E
n,
and for s(x) >
0, lim
n
f
n
(x) = f(x) s(x) > s(x) implies x E
n,
when n is large enough. So

n=1
E
n,
= R and Proposition 3.4.10 (with Proposition 3.4.9) tells us that
_
R
s d =
s
(R) =
s
_

_
n=1
E
n,
_
= lim
n

s
(E
n,
) lim
n
_
R
f
n
d
16 201112 Mathematics MA2224
(using (1) at the last step). That gives us

_
R
s d =
_
R
s d lim
n
_
R
f
n
d.
As that is true for each (0, 1), it follows that
_
R
s d lim
n
_
R
f
n
d, and so
_
R
f d = sup
__
R
s d : s simple measureable, s f
_
lim
n
_
R
f
n
d,
completing the proof.
Corollary 3.5.7. If f, g : R [0, ] are (Lebesgue) measurable functions and 0 then
(a)
_
R
f d =
_
R
f d
(b)
_
R
(f +g) d =
_
R
f d +
_
R
g d
Proof. (a) By Proposition 3.5.1, there is a monotone increasing sequence (f
n
)

n=1
of measurable
simple functions that converges pointwise to f. By the Monotone Convergence Theorem
(3.5.6) we know that
_
R
f d = lim
n
_
R
f
n
d. But also (f
n
)

n=1
is a sequence of mea-
surable simple functions that converges pointwise to f and so we also have
_
R
f d =
lim
n
_
R
f
n
d. Now from Proposition 3.4.6 we have

_
R
f d = lim
n

_
R
f
n
d = lim
n
_
R
f
n
d =
_
R
f d
(b) Again by Proposition 3.5.1, there are monotone increasing sequences (f
n
)

n=1
and (g
n
)

n=1
of measurable simple functions with lim
n
f
n
(x) = f(x) and lim
n
g
n
(x) = g(x) for
each x R. By the Monotone Convergence Theorem (3.5.6) we know that
_
R
f d =
lim
n
_
R
f
n
d and
_
R
g d = lim
n
_
R
g
n
d. But also (f
n
+ g
n
)

n=1
is a monotone
increasing sequence of measurable simple functions that converges pointwise to f + g and
so we also have
_
R
(f +g) d = lim
n
_
R
(f
n
+g
n
) d. From Proposition 3.4.6 we have
_
R
(f +g) d = lim
n
_
R
(f
n
+g
n
) d
= lim
n
__
R
f
n
d +
_
R
f
n
d
_
=
_
R
f d +
_
R
g d
Corollary 3.5.8. If f, g : R [0, ] are (Lebesgue) measurable functions, 0 and E L,
then
_
E
(f +g) d =
_
E
f d +
_
E
g d
Lebesgue integral 17
Proof. Apply Corollary 3.5.7 to
E
(f +g) =
E
f +
E
g to get
_
E
(f +g) d =
_
R

E
(f +g) d
=
_
R
(
E
f +
E
g) d
=
_
R

E
f d +
_
R

E
g d
=
_
E
f d +
_
E
g d
3.6 Integrable functions
Denition 3.6.1. If f : R [, ] is measurable, then we say that f is Lebesgue integrable
if
_
R
f
+
d < and
_
R
f

d < .
For integrable f, we dene
_
R
f d =
_
R
f
+
d
_
R
f

d
Lemma 3.6.2. If f : R [, ] is measurable, then f is Lebesgue integrable if and only if
_
R
[f[ d <
Proof. If f is Lebesgue integrable, then [f[ = f
+
+f

is (Lebesgue) measurable and

_
R
[f[ d =
_
R
f
+
d +
_
R
f

d < .
Conversely, if f is Lebesgue integrable and
_
R
[f[ d < , then f
+
[f[ (and we know f
+
is measurable) and so
_
R
f
+
d
_
R
[f[ d < .
Similarly
_
R
f

d < .
Exercise 3.6.3. If f : R [, ] is (Lebesgue) integrable, show that

_
R
f d

_
R
[f[ d
(an integral version of the triangle inequality).
18 201112 Mathematics MA2224
Lemma 3.6.4. If g, h: R [0, ] are (Lebesgue) measurable functions with both
_
R
g d <
and
_
R
hd < , and if
x R : g(x) = x R : h(x) = = ,
then f = g h is integrable and
_
R
f d =
_
R
g d
_
R
hd.
Proof. We know that f(x) = g(x) h(x) is dened everywhere and f is measurable (by Propo-
sition 3.2.12). Also [f(x)[ g(x) +h(x) and so (via Lemma 3.5.5)
_
R
[f[ d
_
R
(g +h) d =
_
R
g d +
_
R
hd < .
So f is integrable.
From f = g h = f
+
f

we have f

+g = f
+
+h and so by Corollary 3.5.7 we get
_
R
f

d +
_
R
g d =
_
R
(f

+g) d =
_
R
(f
+
+h) d =
_
R
f
+
d +
_
R
hd
which gives us
_
R
f d =
_
R
f
+
d
_
R
f

d =
_
R
g d
_
R
hd.
Theorem 3.6.5 (Linearity of the integral). If f, g : R [, ] are (Lebesgue) integrable
functions and R then
(a) f is integrable and
_
R
f d =
_
R
f d
(b) if f(x) + g(x) is dened for each x R, then f + g is integrable and
_
R
(f + g) d =
_
R
f d +
_
R
g d
Proof. (a) If = 0, the result is easy. For ,= 0, f is measurable (by Proposition 3.2.5) and
_
R
[f[ d = [[
_
R
[f[ d < so that f is integrable. If > 0 then (f)
+
= f
+
,
(f)

= f

, which gives (using Corollary 3.5.7 (a))

_
R
f d =
_
R
f
+
d
_
R
f

d
=
_
R
f
+
d
_
R
f

d
=
__
R
f
+
d
_
R
f

d
_
=
_
R
f d
Lebesgue integral 19
If < 0, then (f)
+
= f

= [[f

, (f)

= f
+
= [[f
+
and so f = (f)
+

(f)

= [[f

[[f
+
. Using Corollary 3.5.7 (a) we have
_
R
f d =
_
R
[[f

d
_
R
[[f
+
d
= [[
__
R
f

d
_
R
f
+
d
_
=
__
R
f
+
d
_
R
f

d
_
=
_
R
f d
(b) f + g is measurable (by Proposition 3.2.5). We know
_
R
f
+
d < ,
_
R
f

d < ,
_
R
g
+
d < and
_
R
g

d < . Also
f +g = (f
+
f

) + (g
+
g

) = (f
+
+g
+
) (f

+g

)
and from Lemma 3.6.4 (and Corollary 3.5.7 (b) too) we nd
_
R
(f +g) d =
_
R
(f
+
+g
+
) d
_
R
(f

+g

) d
=
_
R
f
+
d +
_
R
g
+
d
__
R
f

d +
_
R
g

d
_
=
_
R
f d +
_
R
g d
3.7 Functions integrable on subsets
Denition 3.7.1. If X L and f : X [, ] is measurable on X, then we say that f is
integrable on X if the function F : R [, ] given by
F(x) =
_
f(x) if x X
0 if x R X
is integrable and in that case we dene
_
X
f d =
_
X
F d =
_
R

X
F d
Remark 3.7.2. Recall from Proposition 3.2.11 that F must be (Lebesgue) measurable on R.
It is an easy exercise to show that Theorem 3.6.5 remains true for functions integrable on
X L and
_
X
in place of
_
R
.
20 201112 Mathematics MA2224
Example 3.7.3. If X L has (X) < and f : X R is a constant function f(x) = c, then
f is integrable and
_
X
f d =
_
R
c
X
d = c
_
R

X
d = c(X).
Theorem 3.7.4. If f : [a, b] R is continuous on a nite closed interval, then the Lebesgue
integral
_
[a,b]
f d exists and has the same value as the Riemann integral
_
b
a
f(x) dx.
Proof. Since any such f is necessarily bounded, there is M 0 so that [f(x)[ M for a x
b. By considering f(x) +M in place of f, the proof can be reduced to the case where f(x) 0
for all x [a, b]. Thus we assume f(x) 0.
For this proof we use step functions, rather than simple functions. A step function is a linear
combination

n
j=1

E
j
where
j
R and the sets E
j
are intervals. (We will only want nite
intervals here.)
Consider the lower sums for f with respect to the partitions of [a, b] into 2
n
equal subintervals
of length (b a)/2
n
. For 1 i 2
n
, let
L
n,i
= inf
_
f(x) : a +
b a
2
n
(i 1) x a +
b a
2
n
i
_
so that the corresponding lower sum is
L
n
(f) =
2
n

i=1
L
n,i
b a
2
n
We can make a step function s
n
(x) with integral equal to this sum by letting s
n
(x) = 0 for
x / [a, b] and
s
n
(x) = L
n,i
for a +
b a
2
n
(i 1) x < a +
b a
2
n
i (1 i 2
n
)
(we can make s
n
(b) = L
n,2
n to be careful). We have then s
n
(x) f(x) for x [a, b] and
_
R
s
n
d = L
n
(f).
From the way each partition is a renement of the previous one, it follows that
L
n+1,2i
L
n,i
and L
n+1,2i1
L
n,i
.
Thus s
n
(x) s
n+1
(x) for each x R.
From uniform continuity of f we can deduce that, given any > 0, for all n sufciently large
sup
x[a,b]
[f(x) s
n
(x)[ = sup
x[a,b]
f(x) s
n
(x) <
Lebesgue integral 21
and
0 <
_
b
a
f(x) dx L
n
(f) < (b a).
So we deduce that
lim
n
s
n
(x) = F(x) =
_
f(x) if x [a, b]
0 if x / [a, b]
and
lim
n
L
n
(f) =
_
b
a
f(x) dx
But by the monotone convergence theorem (remember we assumed f(x) 0)
lim
n
L
n
(f) = lim
n
_
R
s
n
d =
_
R
F d =
_
[a,b]
f d
which shows that the two integrals are the same.
R. Timoney March 14, 2012