Lecture7 Borel Sets and Lebesgue Measure
Lecture7 Borel Sets and Lebesgue Measure
Lecture7 Borel Sets and Lebesgue Measure
In this lecture, we discuss the case where the sample space is uncountable. This case is more involved than
the case of a countable sample space, mainly because it is often not possible to assign probabilities to all
subsets of . Instead, we are forced to work with a smaller -algebra. We consider assigning a uniform
probability measure on the unit interval.
Consider the experiment of picking a real number at random from = [0, 1], such that every number is
equally likely to be picked. It is quite apparent that a simple strategy of assigning probabilities to singleton
subsets of the sample space gets into difficulties quite quickly. Indeed,
(i) If we assign some positive probability to each elementary outcome, then the probability of an event
with infinitely many elements, such as A = {1, 21 , 13 , }, would become unbounded.
Thus, we need a different approach to assign probabilities when the sample space is uncountable, such as
= [0, 1]. In particular, we need to assign probabilities directly to specific subsets of . Intuitively, we
would like our uniform measure on [0, 1] to possess the following two properties.
(ii) Translational Invariance. That is, if A [0, 1], then for any x , (A x) = (A) where, the set
A x is defined as
However, the following impossibility result asserts that there is no way to consistently define a uniform
measure on all subsets of [0, 1].
Theorem 7.1 (Impossibility Result) There does not exist a definition of a measure (A) for all subsets
of [0, 1] satisfying (i) and (ii).
Therefore, we must compromise, and consider a smaller -algebra that contains certain nice subsets of
the sample space [0, 1]. These nice subsets are the intervals, and the resulting -algebra is called the
Borel -algebra. Before defining Borel sets, we introduce the concept of generating -algebras from a given
collection of subsets.
7-1
7-2 Lecture 7: Borel Sets and Lebesgue Measure
The -algebra generated by a collection of subsets of the sample space is the smallest -algebra that contains
the collection. More formally, we have the following theorem.
Theorem 7.2 Let C be an arbitrary collection of subsets of , then there exists a smallest -algebra, denoted
by (C), that contains all elements of C. That is, if H is any -algebra such that C H, then (C) H.
(C) is called the -algebra generated by C.
Proof: Let {Fi , i I} denote the collection of all -algebras that contain C. Clearly, the T collection
{Fi , i I} is non-empty, since it contains at least the power set, 2 . Consider the intersection Fi . Since
iI
the intersectionTof -algebras results in a -algebra (homework problem!) and the intersection contains C,
it follows that Fi is a -algebra that contains C. Finally, if C H, then H is one of Fi s for some i I.
T iI
Hence Fi is the smallest -algebra generated by C.
iI
Intuitively, we can think of C as being the collection of subsets of which are of interest to us. Then, (C)
is the smallest -algebra containing all the interesting subsets.
We are now ready to define Borel sets.
Definition 7.3
(a) Consider = (0, 1]. Let C0 be the collection of all open intervals in (0, 1]. Then (C0 ) , the - algebra
generated by C0 , is called the Borel - algebra. It is denoted by B ((0, 1]).
(b) An element of B ((0, 1]) is called a Borel-measurable set, or simply a Borel set.
Thus, every open interval in (0, 1] is a Borel set. We next prove that every singleton set in (0, 1] is a Borel
set.
Lemma 7.4 Every singleton set {b}, 0 < b 1, is a Borel set, i.e., {b} B ((0, 1]).
Proof: Consider the collection of sets set b n1 , b + n1 , n 1 . By the definition of Borel sets,
1 1
b ,b + B ((0, 1]) .
n n
Using the properties of -algebra,
c
1 1
b ,b + B ((0, 1])
n n
c
[ 1 1
= b ,b + B ((0, 1])
n=1
n n
!c
\ 1 1
= b ,b + B ((0, 1])
n=1
n n
\ 1 1
= b ,b + B ((0, 1]) . (7.1)
n=1
n n
Lecture 7: Borel Sets and Lebesgue Measure 7-3
Non-Borel sets
Banach-Tarski paradox (a bizzare phenomenon about cutting up the surface of a sphere. See https:
//www.youtube.com/watch?v=Tk4ubu7BlSk
The cardinality of the Borel -algebra (on the unit interval) is the same as the cardinality of the
reals. Thus, the Borel -algebra is a much smaller collection than the power set 2[0,1] . See https:
//math.dartmouth.edu/archive/m103f08/public_html/borel-sets-soln.pdf
In this section, we discuss a formal procedure to define a probability measure on a general measurable space
(, F). Specifying the probability measure for all the elements of F directly is difficult, so we start with a
smaller collection F0 of interesting subsets of , which need not be a -algebra. We should take F0 to be
rich enough, so that the -algebra it generates is same as F. Then we define a function P0 : F0 [0, 1],
such that it corresponds to the probabilities we would like to assign to the interesting subsets in F0 . Under
certain conditions, this function P0 can be extended to a legitimate probability measure on (, F) by using
the following fundamental theorem from measure theory.
Theorem 7.5 (Caratheodorys extension theorem) Let F0 be an algebra of subsets of , and let F =
(F0 ) be the -algebra that it generates. Suppose that P0 is a mapping from F0 to [0, 1] that satisfies
P0 () = 1, as well as countable additivity on F0 .
Then, P0 can be extended uniquely to a probability measure on (, F). That is, there exists a unique proba-
bility measure P on (, F) such that P (A) = P0 (A) for all A F0 .
7-4 Lecture 7: Borel Sets and Lebesgue Measure
Consider = (0, 1]. Let F0 consist of the empty set and all sets that are finite unions of the intervals of the
form (a, b]. A typical element of this set is of the form
Lemma 7.6
a) F0 is an algebra
b) F0 is not a -algebra
c) (F0 ) = B
Proof:
c) First, the null set is clearly a Borel set. Next, we have already seen that every interval of the form
(a, b] is a Borel set. Hence, every element of F0 (other than the null set), which is a finite union of
such intervals, is also a Borel set. Therefore, F0 B. This implies (F0 ) B.
Next we show that B (F0 ). For any interval of the form (a, b) in C0 , we can write (a, b) =
a, b n1 . Since every interval of the form a, b n1 F0 , a countable number of unions of
S
n=1
such intervals belongs to (F0 ). Therefore, (a, b) (F0 ) and consequently, C0 (F0 ). This gives
(C0 ) (F0 ). Using the fact that (C0 ) = B proves the required result.
n
P
P0 () = 0 and P0 (F ) = (bi ai ).
i=1
Lecture 7: Borel Sets and Lebesgue Measure 7-5
Note that P0 () = P0 ((0, 1]) = 1. Also, if (a1 , b1 ] , (a2 , b2 ] , . . . , (an , bn ] are disjoint sets, then
n
! n
[ X
P0 ((ai , bi ]) = P0 ((ai , bi ])
i=1 i=1
n
X
= (bi ai )
i=1
The Lebesgue measure formalizes the notion of length. This suggests that the Lebesgue measure of a singleton
should be zero. This can be shown as follows. Let b (0, 1]. Using (7.2), we write
!
\ 1
P ({b}) = P b ,b
n=1
n
1
P (An ) = (7.3)
n
Since An is a decreasing sequence of nested sets,
!
\
P ({b}) =P An
n=1
= lim P (An )
n
1
= lim
n n
=0
where the second equality follows from the continuity of probability measures.
Since any countable set is a countable union of singletons, the probability of a countable set is zero. For
example, under the uniform measure on (0, 1], the probability of the set of rationals is zero, since the rational
numbers in (0, 1] form a countable set.
For = (0, 1], the Lebesgue measure is also a probability measure. For other intervals (for example =
(0, 2]), it will only be a finite measure, which can be normalized as appropriate to obtain a uniform probability
measure.
Definition 7.7 Let (, F, P) be a probability space. An event A is said to occur almost surely (a.s) if
P(A) = 1.
Lebesgue Measure of the Cantor set: Consider the cantor set K. It is created by repeatedly removing
the open middle thirds of a set of line segments. Consider its complement. It contains countable number of
disjoint intervals. Hence we have:
1
1 2 4
P(K c ) = + + + = 3 2 = 1.
3 9 27 1 3
Therefore P(K) = 0. It is very interesting to note that though the Cantor set is equicardinal with (0, 1], its
Lebesgue measure is equal to 0 while the Lebesgue measure of (0, 1] is equal to 1.
We now extend the definition of Lebesgue measure on [0, 1] to the real line, R. We first look at the definition
of a Borel set on R. This can be done in several ways, as shown below.
Let C be a collection of open intervals in R. Then B(R) = (C) is the Borel set on R.
Let D be a collection of semi-infinite intervals {(, x]; x R}, then (D) = B(R).
A R is said to be a Borel set on R, if A (n, n + 1] is a Borel set on (n, n + 1] n Z.
Exercise: Verify that the three statements are equivalent definitions of Borel sets on R.
(R) =
() = 0
The countable additivity property
We see that
Pn (R (n, n + 1]) = 1, n I
Hence we have
X
(R) = 1=
n=
Now consider (n, n + 1]. This is a null set for all n. Hence we have,
Pn ( (n, n + 1]) = 0, n I
which implies,
X
() = Pn ( (n, n + 1]) = 0
n=
Lecture 7: Borel Sets and Lebesgue Measure 7-7
We now need to prove the countable additivity property. For this we consider Ai B(R) such that the
sequence A1 , A2 , . . . , An , . . . are arbitrary pairwise disjoint sets in B(R). Therefore we obtain,
[
X
[
( Ai ) = Pn ( Ai (n, n + 1])
i=1 n= i=1
X X
= Pn (Ai (n, n + 1])
n= i=1
X X
= Pn (Ai (n, n + 1])
i=1 n=
The second equality above comes from the fact that the probability measure has countable additivity prop-
erty. The last equality above comes from the fact that the summations can be interchanged (from Fubinis
theorem). We also have the following:
X
(Ai ) = Pn (Ai (n, n + 1])
n=
7.5 Exercises
1. Let F be a -algebra corresponding to a sample space . Let H be a subset of that does not belong
to F. Consider the collection G of all sets of the form (H A) (H c B), where A and B F.
6. Let = [0, 1], and let F3 consist of all countable subsets of , and all subsets of having a countable
complement. It can be shown that F3 is a -algebra (Refer Lecture 4, Exercises, 6(d)). Let us define
P(A) = 0 if A countable, and P(A) = 1 if A has a countable complement. Is (, F3 , P) a legitimate
probability space?
7. We have seen in 4(c) that the Cantor set is Borel-measurable. Show that the Cantor set has zero
Lebesgue measure. Thus, although the Cantor set can be put into a bijection with [0, 1], it has zero
Lebesgue measure!
References
[1] Rosenthal, J. S. (2006). A first look at rigorous probability theory (Vol. 2). Singapore: World Scientific.
[2] Williams, D. (1991). Probability with martingales. Cambridge university press.