MTH309 (2025-26, E VEN S EMESTER )
Problem Set 2: σ-fields and measures
1. For any non-empty D ⊆ R, the restriction of the Borel σ-field B to the set D is called
the Borel σ-field on D and often also written as B( D ). Show that the Borel σ-field on an
interval I is generated by the class {( a, b) ⊆ I | a, b are rational } and also by the class
{( a, b] ⊆ I | a, b are rational }.
2. What is the Borel σ-field on the set of (i) all integers? (ii) all rationals? (iii) all irrationals?
3. Let I1 and I2 be any two intervals. Show that the Borel σ-field on D = I1 ∪ I2 consists
precisely of sets of the form A1 ∪ A2 , where A1 and A2 are Borel subsets of I1 and I2
respectively.
4. For d > 1, one can define, as in (1), the Borel σ-field on any non-empty D ⊆ Rk . Show
that the Borel σ-field on the set D = Rd−1 × {0} consists precisely of the sets B × {0} with
B ∈ B d−1 , and, that the Borel σ-field on D = {( x, x ) | x ∈ R} consists precisely of sets of
the form {( x, x ) | x ∈ B}, B ∈ B .
5. On R2 , consider the class of all “vertical cylinders” of the type ( a, b) × R, where a < b are
real numbers. Show that the σ-field generated by this class consists precisely of sets of the
form B × R with B varying over all Borel subsets of R. What will be the σ-field generated
by the “horizontal cylinders”? What is the intersection of this two σ-fields?
6. Suppose Ω is a countably infinite set and F consists of all subsets of Ω. Define µ( A) = 0 if
A is finite and µ( A) = ∞ if A is infinite.
(a) Show that µ is finitely additive but not countably additive.
1
(b) Show that Ω is a limit of an increasing sequence of sets An where µ( An ) = 0 for all n
but µ(Ω) = ∞.
7. Let Ω = {1, 2, . . .} and let G be the field consisting of all finite subsets of Ω and their
complements. If A is finite, define µ( A) = 0, and if Ac is finite, set µ( A) = 1
(a) Show that µ is finitely additive but not countably additive on G .
(b) Show that Ω is a limit of an increasing sequence of sets An where µ( An ) = 0 for all n
but µ(Ω) = 1.
8. Let F be the field of finite disjoint unions of right semi-closed intervals of R. Define the set
function µ as follows
µ((−∞, a]) = a where a ∈ R,
µ(( a, b]) = b − a where a, b ∈ R and a < b,
µ((b, ∞)) = −b where b ∈ R,
µ(R) = 0,
n
µ(∪in=1 Ai ) = ∑ µ ( Ai )
i =1
if A1 , A2 , . . . , An are disjoint right semi-closed intervals of R
(a) Show that µ is finitely additive,
(b) Show that µ is finite but unbounded on F.
9. (Bonferroni’s inequality for two events) Let A, B be two events in a probability space
(Ω, F , P). Show that
P( A ∩ B) ≥ max{0, P( A) + P( B) − 1}.
Can you generalize this to any finite number of events?
10. If µ is non-negative and finitely additive set functions on a field F , then show that for
disjoint A1 , A2 , . . . sets in F with ∪n An ∈ F , we have
µ(∪n An ) ≥ ∑ µ ( A n ).
n
2
11. (Inclusion-Exclusion principle for finitely many sets) Let (Ω, F , P) be a probability space
and let A1 , . . . , An be events. Show that
P(∪in=1 Ai ) = S1,n − S2,n + S3,n − · · · + (−1)n−1 Sn,n
where
Sk,n := ∑ P ( A i1 ∩ · · · ∩ A i k . )
1≤i1 <i2 <···<ik ≤n
12. Let P be a probability measure on a σ-field F of subsets of a set Ω. Let A1 , A2 , . . . be a
sequence in F , then show that
(a) P(lim inf An ) ≤ lim inf P( An ) ≤ lim sup P( An ) ≤ P(lim sup An ),
(b) If lim An = A then lim P( An ) = P( A).
13. Show that if µ is a finite measure, there cannot be uncountably many disjoint sets A such
that µ( A) > 0.
14. Let (Ω, F , P) be a probability space and An , Bn ∈ F with Bn ⊆ An for all n ∈ N. Prove or
disprove:
∞
P(∪∞ ∞
n=1 An ) − P(∪n=1 Bn ) ≤ ∑ [P( An ) − P( Bn )].
n =1
15. Let (Ω, F , P) be a probability space.
(a) Suppose { En }n is a sequence of events in F with P(∪n En ) = ∑n P( En ). Show that
P( Ei ∩ Ej ) = 0 for all i ̸= j.
(b) Prove or disprove: Suppose { En }n is a sequence of events in F with P( En ) = 1 for all
n. Then P(∩n En ) = 1.
16. Let (Ω, F , P) be a probability space. Let G = { A ∈ F | P( A) = 0 or 1}. Prove or disprove
that G is a σ-field.