Department of Mathematics, IIT Roorkee

Autumn Semester 2025–26

MA 533 - Measure Theory
Assignment 2
Course Instructor: Dr. Manil T. Mohan
S
(1) Show that S = n∈K (n, n + 1] : K ⊂ Z is a σ-algebra on R.
(2) (a) Suppose S is the smallest σ-algebra on R containing {(r, s] : r, s ∈ Q}. Prove that S
is the collection of Borel subsets of R.
(b) Suppose S is the smallest σ-algebra on R containing {[r, ∞) : r ∈ Q}. Prove that S
is the collection of Borel subsets of R.
(3) Prove that the collection of Borel subsets of R is translation invariant. More precisely,
prove that if B ⊆ R is a Borel set and t ∈ R, then t + B is a Borel set.
(4) Give an example of a measurable space (X, S) and a function f : X → R such that |f | is
S-measurable but f is not S-measurable.
(5) Suppose T is a σ-algebra on a set Y and X ∈ T . Let S = {E ∈ T : E ⊆ X}.
(a) Show that S = {F ∩ X : F ∈ T }.
(b) Show that S is a σ-algebra on X.
(6) Suppose f : R → R is a function.
(a) For k ∈ Z+ , let
 
1
Gk = a ∈ R : there exists δ > 0 such that |f (b) − f (c)| < for all b, c ∈ (a − δ, a + δ) .
k
Prove that Gk is an open subset of R for each k ∈ Z+ .
(b) Prove that the set of points at which f is continuous equals ∞
T
k=1 Gk .
(c) Conclude that the set of points at which f is continuous is a Borel set.
(7) Suppose (X, S) is a measurable space, E1 , . . . , En are disjoint subsets of X, and c1 , . . . , cn
are distinct nonzero real numbers. Prove that c1 χE1 + · · · + cn χEn is an S-measurable
function if and only if E1 , . . . , En ∈ S.
(8) (a) Suppose f1 , f2 , . . . is a sequence of functions from a set X to R. Explain why
∞ [ ∞ \ ∞  
\
−1 1 1
{x ∈ X : the sequence f1 (x), f2 (x), . . . has a limit in R} = (fj − fk ) − , .
n=1 j=1 k=j
n n

(b) Suppose (X, S) is a measurable space and f1 , f2 , . . . is a sequence of S-measurable

functions from X to R. Prove that
{x ∈ X : the sequence f1 (x), f2 (x), . . . has a limit in R}
is an S-measurable subset of X.
(9) Suppose X is a set and E1 , E2 , . . . is a disjoint sequence of subsets of X such that ∞
S
k=1 Ek =
X. Let ( )
[
S= Ek : K ⊆ Z+ .
k∈K
(a) Show that S is a σ-algebra on X.
(b) Prove that a function from X to R is S-measurable if and only if the function is
constant on Ek for every k ∈ Z+ .
(10) Suppose S is a σ-algebra on a set X and A ⊆ X. Let
SA := {E ∈ S : A ⊆ E or A ∩ E = ∅}.
1
2

(a) Prove that SA is a σ-algebra on X.

(b) Suppose f : X → R is a function. Prove that f is measurable with respect to SA if
and only if f is measurable with respect to S and f is constant on A.
(11) Suppose X is a Borel subset of R and f : X → R is a function such that {x ∈ X :
f is not continuous at x} is a countable set. Prove f is a Borel measurable function.
(12) Suppose f : R → R is differentiable at every element of R. Prove that f 0 is a Borel
measurable function from R to R.
(13) Suppose X is a nonempty set and S is the σ-algebra on X consisting of all subsets of X
that are either countable or have a countable complement in X. Give a characterization
of the S-measurable real-valued functions on X.
(14) (a) Suppose B ⊆ R and f : B → R is an increasing function. Prove that f is continuous
at every element of B except for a countable subset of B.
(b) Suppose f : R → R is a strictly increasing function. Prove that the inverse function
f −1 : f (R) → R is a continuous function. [Note that this exercise does not have as a
hypothesis that f is continuous.]
(c) Suppose f : R → R is a strictly increasing function and B ⊆ R is a Borel set. Prove
that f (B) is a Borel set.
(15) Prove or give a counterexample: If (X, S) is a measurable space and f : X → [−∞, ∞] is
a function such that f −1 ((a, ∞)) ∈ S for every a ∈ R, then f is an S-measurable function.

S
Proof. We need to verify that S = n∈K (n, n + 1] : K ⊂ Z satisfies the three axioms of a σ-
algebra on R.

1. Contains the Whole Space. Take K = Z. Then:

[
(n, n + 1] = R ∈ S
n∈Z

since every real number x belongs to some interval (n, n + 1] where n = bxc.
S
2. Closed Under Complementation. Let A = n∈K (n, n + 1] ∈ S. Its complement is:
[
Ac = R \ A = (n, n + 1]
n∈Z\K

This union is in S because Z \ K ⊂ Z.

3. Closed Under Countable Unions. Let {Ai }∞

S
i=1 ⊂ S where each Ai = n∈Ki (n, n + 1] for
some Ki ⊂ Z. Then:
∞
[ ∞ [
[ [
Ai = (n, n + 1] = (n, n + 1]
i=1 i=1 n∈Ki n∈ ∞
S
i=1 Ki
S∞
Since i=1 Ki ⊂ Z, the countable union remains in S.

Conclusion. S satisfies all three axioms of a σ-algebra:

(1) R ∈ S
(2) Closed under complementation
(3) Closed under countable unions
Therefore, S is indeed a σ-algebra on R. 
S is a σ-algebra on R as it satisfies all three defining properties.
 3

Problem (a)
Suppose S is the smallest σ-algebra on R containing {(r, s] : r, s ∈ Q}. Prove that S is the
collection of Borel subsets of R.

Solution. We need to show that S = B(R), where B(R) is the Borel σ-algebra on R.

Step 1: S ⊆ B(R).
• The Borel σ-algebra B(R) contains all open intervals, and hence all half-open intervals
(a, b], because:
∞  
\ 1
(a, b] = a, b + .
n=1
n
1


Since a, b + n is open, (a, b] is a countable intersection of open sets, hence Borel.
• Specifically, for rational r, s, (r, s] is Borel.
• S is the smallest σ-algebra containing all (r, s], and since B(R) is a σ-algebra containing
these sets, S ⊆ B(R).

Step 2: B(R) ⊆ S.
• Any open set U ⊆ R can be written as a countable union of open intervals (an , bn ) because
R is second-countable.
• We can approximate (a, b) using (r, s]:
∞
[
(a, b) = (an , bn ],
n=1

where an is a sequence of rationals decreasing to a and bn is a sequence of rationals in-

creasing to b.
• Since (an , bn ] ∈ S because an , bn ∈ Q, (a, b) is a countable union of sets in S, so (a, b) ∈ S.
• Thus, S contains all open sets, and since B(R) is the smallest σ-algebra containing all open
sets, B(R) ⊆ S.

Conclusion. Since both inclusions hold, S = B(R).

S is the collection of Borel subsets of R because it is the smallest σ-algebra containing all open
intervals, which generate the Borel σ-algebra.

Problem (b)
Suppose S is the smallest σ-algebra on R containing {[r, ∞) : r ∈ Q}. Prove that S is the
collection of Borel subsets of R.

Solution. Again, we show S = B(R).

Step 1: S ⊆ B(R).
• The sets [r, ∞) are closed in R, hence Borel sets because the Borel σ-algebra contains all
closed sets.
• S is the smallest σ-algebra containing these sets, and since B(R) is a σ-algebra that contains
them, S ⊆ B(R).
4

Step 2: B(R) ⊆ S.
• The complement of [r, ∞) is (−∞, r), which is in S because σ-algebras are closed under
complements.
• For any rational r, s with r < s, consider:
(−∞, s) ∩ [r, ∞) = [r, s).
Since S is closed under finite intersections, [r, s) ∈ S.
• To get (r, s), note that:
∞  
[ 1
(r, s) = r + ,s .
n=1
n
Since r + n1 , s ∈ S, and S is closed under countable unions, (r, s) ∈ S.


• For any real numbers a, b, find sequences of rationals rn ↓ a and sn ↑ b such that:
∞
[
(a, b) = (rn , sn ).
n=1

Since (rn , sn ) ∈ S, (a, b) ∈ S.

• Any open set U ⊆ R is a countable union of such open intervals (an , bn ), hence U ∈ S.
• Therefore, S contains all open sets, and B(R) ⊆ S.
Conclusion. Both inclusions hold, so S = B(R).
S is the collection of Borel subsets of R because it is the smallest σ-algebra containing all
open intervals, which can be generated from the sets [r, ∞) for r ∈ Q.
Theorem 0.1. If f : R → R is a homeomorphism, then f (B) = B, where B is the Borel σ-algebra
on R.
Proof. We prove both inclusions:
1. f (B) ⊆ B.
• Let A ∈ B. Since f is continuous, f −1 pulls back open sets to open sets.
• The Borel σ-algebra B is generated by open sets, so for any Borel set A, f (A) = (f −1 )−1 (A)
is Borel because f −1 is continuous and thus measurable.
2. B ⊆ f (B).
• Since f is a homeomorphism, f −1 is also a homeomorphism.
• By the same argument as above, for any B ∈ B, f −1 (B) ∈ B.
• Thus B = f (f −1 (B)) ∈ f (B) because f −1 (B) ∈ B.
Conclusion. We’ve shown both inclusions, therefore f (B) = B. 
For any homeomorphism f : R → R, f (B) = B where B is the Borel σ-algebra on R.
Example 0.2. Consider the measurable space (R, S) where S is the σ-algebra of Lebesgue mea-
surable sets, and let A ⊂ R be a non-measurable set (e.g., a Vitali set). Define the function:
(
1 if x ∈ A
f (x) =
−1 if x ∈ /A
Then:
(1) |f | is S-measurable
(2) f is not S-measurable
 5

Proof. 1. |f | is measurable
The absolute value function is:
|f |(x) = 1 for all x ∈ R
This is a constant function, and all constant functions are measurable with respect to any σ-
algebra. Specifically, for any Borel set B ⊆ R:
(
∅ if 1 ∈
/B
|f |−1 (B) =
R if 1 ∈ B

Both ∅ and R are in S.

2. f is not measurable
Consider the preimage of {1}:
f −1 ({1}) = A
Since A was chosen to be non-measurable, and {1} is a Borel set, this shows f is not S-measurable.

Summary: The function f that takes value 1 on a non-measurable set A and −1 elsewhere has:
• |f | ≡ 1 is trivially measurable
• f itself is non-measurable because f −1 ({1}) = A ∈
/S
This provides the required example.

Problem Statement
Suppose T is a σ-algebra on a set Y and X ∈ T . Let S = {E ∈ T : E ⊆ X}.
(1) Show that S = {F ∩ X : F ∈ T }.
(2) Show that S is a σ-algebra on X.

Solution
Part (a): Equality of Collections. We need to show:
S = {F ∩ X : F ∈ T }

Proof. We prove both inclusions:

1. S ⊆ {F ∩ X : F ∈ T }. For any E ∈ S, by definition:

• E∈T
• E⊆X
We can write E = E ∩ X (since E ⊆ X). Thus, E is of the form F ∩ X where F = E ∈ T .

2. {F ∩ X : F ∈ T } ⊆ S. For any F ∈ T :
• F ∩ X ∈ T (since T is closed under intersections and X ∈ T )
• F ∩X ⊆X
Thus, F ∩ X ∈ S by definition.

Conclusion. Both inclusions hold, so the sets are equal. 

6

Part (b): S is a σ-algebra on X. We verify the three properties of a σ-algebra:

Proof. 1. X ∈ S
• X ∈ T (given)
• X ⊆ X (trivially)
Thus, X ∈ S.
2. Closed under complementation (relative to X)
For E ∈ S:
• E ∈ T and E ⊆ X
• The complement in X is X \ E = X ∩ E c
• E c ∈ T (since T is a σ-algebra)
• Thus, X \ E = X ∩ E c ∈ S by part (a)
3. Closed under countable unions
Let {En }∞n=1 ⊆ S:
• S
Each En ∈ T and En ⊆ X
• S∞ n=1 En ∈ T (since T is a σ-algebra)
• ∞ n=1 ESn ∞⊆ X
• Thus, n=1 En ∈ S
Conclusion
S contains X and is closed under complementation and countable unions, hence it is a σ-algebra
on X. 
(1) For k ∈ Z+ , let
 
1
Gk = a ∈ R : there exists δ > 0 such that |f (b) − f (c)| < for all b, c ∈ (a − δ, a + δ) .
k
Prove that Gk is an open subset of R for each k ∈ Z+ .
Proof. Let a ∈ Gk . By definition, there exists δ > 0 such that for all b, c ∈ (a − δ, a + δ),
we have |f (b) − f (c)| < k1 .
Consider any x ∈ (a − δ/2, a + δ/2). Let δ 0 = δ/2 − |x − a| > 0. Then for any
y, z ∈ (x − δ 0 , x + δ 0 ), we have:
y, z ∈ (a − δ, a + δ)
because:
|y − a| ≤ |y − x| + |x − a| < δ 0 + |x − a| = δ/2
and similarly for z. Therefore:
1
|f (y) − f (z)| <
k
This shows that x ∈ Gk with the same δ working for all points in the neighborhood of a.
Hence, (a − δ/2, a + δ/2) ⊆ Gk , proving Gk is open. 
T∞
(2) Prove that the set of points at which f is continuous equals k=1 Gk .
Proof. Let C be the set of continuity points of f .
(⊆) If a ∈ C, then for any k ∈ Z+ , there exists δ > 0 such that:
1
|x − a| < δ =⇒ |f (x) − f (a)| <
2k
 7

Then for any b, c ∈ (a − δ, a + δ):

1 1 1
|f (b) − f (c)| ≤ |f (b) − f (a)| + |f (a) − f (c)| < + =
2k 2k k
T∞
Thus a ∈ Gk T for all k, so a ∈ k=1 Gk .
(⊇) If a ∈ ∞ 1
k=1 Gk , then for any  > 0, choose k with k < . Since a ∈ Gk , there exists
δ > 0 such that:
1
|b − a| < δ =⇒ |f (b) − f (a)| ≤ sup |f (c) − f (a)| ≤ < 
c∈(a−δ,a+δ) k
Thus f is continuous at a. 
(3) Conclude that the set of points at which f is continuous is a Borel set.
Proof. From part (a), each Gk is open. From part (b), the continuity set is:
∞
\
C= Gk
k=1

Since each Gk is open (and hence Borel), and the Borel σ-algebra is closed under countable
intersections, C is a Borel set. 
Theorem 0.3. Let (X, S) be a measurable space, E1 , . . . , En be disjoint subsets of X, and c1 , . . . , cn
be distinct nonzero real numbers. Then the function
n
X
f= ci χEi
i=1

is S-measurable if and only if Ei ∈ S for all i = 1, . . . , n.

Proof. We prove both directions of the equivalence.

(⇒) Assume f is S-measurable. For each ci , since f is measurable and {ci } is a Borel set (as
a singleton in R), we have:
f −1 ({ci }) ∈ S
Now observe that:
f −1 ({ci }) = Ei
because:
• If x ∈ Ei , then f (x) = ci (since the Ej ’s are disjoint and the cj ’s are distinct)
• If f (x) = ci , then x must be in Ei (again by disjointness and distinctness)
Thus Ei = f −1 ({ci }) ∈ S for each i.

(⇐) Assume each Ei ∈ S. We need to show f is measurable. For any Borel set B ⊆ R, consider:
[
f −1 (B) = Ei
{i:ci ∈B}

This holds because:

• If x ∈ Ei and ci ∈ B, then f (x) = ci ∈ B, so x ∈ f −1 (B)
• Conversely, if f (x) ∈ B, then x ∈ Ei for some i with ci = f (x) ∈ B
Since each Ei ∈ S and S is closed under finite unions, f −1 (B) ∈ S. Thus f is S-measurable.
8

Special Case Handling. Note that:

n
[
−1
f ({0}) = X \ Ei
i=1

which is also in S when all Ei ∈ S, but we don’t need to consider this case explicitly since we
6 0.
assumed all ci = 

Problem 1: Measurable Simple Functions

Theorem 0.4. Let (X, S) be a measurable space, E1 , . . . , En be disjoint subsets of X, and c1 , . . . , cn
be distinct nonzero real numbers. Then the function
n
X
f= ci χEi
i=1

is S-measurable if and only if Ei ∈ S for all i = 1, . . . , n.

Proof. (⇒) Suppose f is S-measurable. For each i, since {ci } is a Borel set and f is measurable,
we have:
Ei = f −1 ({ci }) ∈ S
This equality holds because the Ei are disjoint and the ci are distinct.
(⇐) Suppose each Ei ∈ S. For any Borel set B ⊆ R:
[
f −1 (B) = Ei ∈ S
{i|ci ∈B}

since S is closed under finite unions. Thus f is measurable. 

Problem 2: Measurability of Limit Points

(1) For a sequence of functions f1 , f2 , . . . : X → R, the set where the sequence converges is:
∞ [
\ ∞ \
∞
(fj − fk )−1 − n1 , n1



n=1 j=1 k=j

Proof. A sequence (fk (x))∞

k=1 converges iff it is Cauchy, meaning:
1
∀n ∈ N, ∃j ∈ N such that ∀k ≥ j, |fj (x) − fk (x)| < n

This is exactly equivalent to:

∞ [
\ ∞ \
∞
x∈ {x : |fj (x) − fk (x)| < n1 }
n=1 j=1 k=j

which gives the stated expression. 

(2) If (X, S) is a measurable space and each fk is S-measurable, then the convergence set is
S-measurable.
Proof. Each (fj − fk )−1 ((− n1 , n1 )) ∈ S because:
• fj − fk is measurable (difference of measurable functions)
• (− n1 , n1 ) is a Borel set
 9

The set is built using countable intersections and unions:

\∞
(fj − fk )−1 ((− n1 , n1 )) ∈ S (countable intersection)
k=j
[∞ \∞
··· ∈ S (countable union)
j=1 k=j
∞
\
··· ∈ S (countable intersection)
n=1
Thus the entire expression is S-measurable. 

Problem Statement
S∞
Let X be a set with a disjoint sequence {Ek }∞
k=1 of subsets such that k=1 Ek = X. Define
( )
[
S= Ek : K ⊆ Z+ .
k∈K

(1) Show that S is a σ-algebra on X.

(2) Prove that f : X → R is S-measurable if and only if f is constant on each Ek .

Solution
Part (a): S is a σ-algebra.
Proof. We verify the three properties of aSσ-algebra:
1. Contains X: Take K = Z+ . Then k∈Z+ Ek = X ∈ S.
S
2. Closed under complements: For any A = k∈K Ek ∈ S, its complement is
[
Ac = Ek ∈ S
k∈Z+ \K

since Z+ \ K ⊆ Z+ .
3. Closed under countable unions: Let {An }∞
S
n=1 ⊆ S where An = k∈Kn Ek . Then
∞
[ ∞ [
[ [
An = Ek = Ek ∈ S
S∞
n=1 n=1 k∈Kn k∈ n=1 Kn

because ∞
S
n=1 Kn ⊆ Z+ .
Thus S satisfies all σ-algebra axioms. 
Part (b): Measurability Condition.
Proof. (⇒) Suppose f is S-measurable. For any k ∈ Z+ and any x, y ∈ Ek , we must show
f (x) = f (y).
Assume for contradiction that f (x) 6= f (y). Then there exists an open interval I containing
f (x) but not f (y). Since f is measurable, f −1 (I) ∈ S.
But Ek ∩f −1 (I) contains x but not y, so f −1 (I) cannot be a union of whole Ej sets, contradicting
the definition of S.
(⇐) Suppose f is constant on each Ek . Let f |Ek = ck . For any Borel set B ⊆ R,
[
f −1 (B) = Ek ∈ S
{k|ck ∈B}
10

since it’s a union of the basic sets Ek . Thus f is S-measurable. 

Problem Statement
Let S be a σ-algebra on X and A ⊆ X. Define
SA := {E ∈ S : A ⊆ E or A ∩ E = ∅}.
(1) Prove that SA is a σ-algebra on X.
(2) Prove that f : X → R is SA -measurable iff f is S-measurable and constant on A.

Solution
Part (a): SA is a σ-algebra.
Proof. We verify the three σ-algebra axioms:
1. Contains X: Since X ∈ S and A ⊆ X, we have X ∈ SA .
2. Closed under complements: For E ∈ SA , consider two cases:
• If A ⊆ E, then A ∩ E c = ∅ (since E c = X \ E)
• If A ∩ E = ∅, then A ⊆ E c
In both cases, E c ∈ SA .
3. Closed under countable unions: Let {En } ⊆ SA . There are two possibilities:
S
• All En satisfy A ∩ En = ∅: Then A ∩ (S En ) = ∅
• Some Em satisfies A ⊆ Em : Then A ⊆ En
S
In both cases, En ∈ SA . 
Part (b): Measurability Condition.
Proof. (⇒) Suppose f is SA -measurable.
1. S-measurability: Since SA ⊆ S, any f −1 (B) ∈ SA implies f −1 (B) ∈ S.
2. Constant on A: Suppose for contradiction f takes two values a 6= b on A. Take disjoint
open intervals Ia 3 a and Ib 3 b. Then:
• f −1 (Ia ) ∩ A 6= ∅ (contains points mapping to a)
• f −1 (Ia ) 6⊇ A (since b ∈
/ Ia )
−1
This contradicts f (Ia ) ∈ SA .
(⇐) Suppose f is S-measurable and constant on A (say f |A ≡ c). For any Borel B ⊆ R:
• If c ∈ B, then A ⊆ f −1 (B)
• If c ∈/ B, then A ∩ f −1 (B) = ∅
Since f −1 (B) ∈ S by S-measurability, we have f −1 (B) ∈ SA . 
Theorem 0.5. Let X be a Borel subset of R and f : X → R a function such that the set
of discontinuity points D = {x ∈ X : f is not continuous at x} is countable. Then f is Borel
measurable.
Proof. We will prove this through several steps:
Step 1: Structure of Discontinuity Set. By hypothesis, D is countable. Since:
• X is Borel (given)
• All countable sets are Borel (as they are countable unions of singletons, and singletons are
closed)
both D and X \ D are Borel sets.
 11

Step 2: Restriction to Continuity Points. On X \ D, f is continuous. Since:

• The restriction of a continuous function to a subset is continuous
• Continuous functions are Borel measurable (preimages of open sets are open in the sub-
space, hence Borel)
f |X\D is Borel measurable.
Step 3: Handling Discontinuity Points. On D, we have:
• D is countable, say D = {d1 , d2 , . . .}
• For each di , {di } is Borel (as it’s closed)
• f |D is measurable because for any Borel B ⊆ R:
[
f |−1
D (B) = {di }
{i:f (di )∈B}

which is a countable union of Borel sets

Step 4: Combining Both Parts. We can write f as:
(
f |X\D (x) x ∈ X \ D
f (x) =
f |D (x) x∈D
Both pieces are Borel measurable, and since X \ D and D are Borel sets, their union is Borel.
Therefore, f is Borel measurable.
Conclusion. The function f is Borel measurable because:
(1) It is continuous (hence Borel measurable) on X \ D
(2) It is trivially Borel measurable on the countable set D
(3) The domain can be partitioned into these two Borel sets

Problem. Suppose X is a Borel subset of R, and f : X → R is a function such that
Df := {x ∈ X : f is not continuous at x}
is a countable set. Prove that f is a Borel measurable function.

Solution.
Let us denote the set of continuity points of f by
C := X \ Df = {x ∈ X : f is continuous at x}.
Since Df is countable and X is a Borel set, both Df and C are Borel subsets of R.
Define the restrictions:
fC := f |C : C → R, fDf := f |Df : Df → R.
Then, the full function f is given by
f = fC ∪ fDf .
Step 1: Since fC is continuous on the Borel set C, it is Borel measurable.
Step 2: The function fDf is defined on a countable set Df , and any function from a countable
set into R is Borel measurable, because every subset of a countable set is Borel measurable.
Step 3: For any Borel set B ⊂ R,
f −1 (B) = fC−1 (B) ∪ fD−1f (B),
where both fC−1 (B) ⊂ C and fD−1f (B) ⊂ Df are Borel subsets of X.
12

Thus, f −1 (B) is a Borel subset of X, and hence f is Borel measurable.

Conclusion: f is a Borel measurable function.

Theorem 0.6. If f : R → R is differentiable everywhere, then its derivative f 0 is Borel measurable.
Proof. We will construct f 0 as a limit of measurable functions.
Step 1: Define Difference Quotients. For each n ∈ N, define:
fn (x) = n f x + n1 − f (x)



These are difference quotients approximating the derivative.

Step 2: Measurability of fn . Each fn is measurable because:
• f is differentiable, hence continuous, hence measurable
• x 7→ x + n1 is measurable (continuous)
• Compositions and linear combinations of measurable functions are measurable
Step 3: Pointwise Convergence. By differentiability of f :
f 0 (x) = lim fn (x) for all x ∈ R
n→∞

Step 4: Limit of Measurable Functions. The Borel σ-algebra is closed under pointwise limits.
Since:
• Each fn is Borel measurable
• f 0 is the pointwise limit of the fn sequence
we conclude f 0 is Borel measurable.
Alternative Approach via Continuity Points. We can also observe that:
• The derivative f 0 is the limit of continuous functions (difference quotients)
• The set of discontinuities of f 0 (if any) must be Fσ countable union of closed sets)
• Any function whose discontinuity set is Fσ is Borel measurable

Definition 0.7. Let X be a nonempty set and S the σ-algebra consisting of all subsets of X that
are either countable or have countable complement in X.
Theorem 0.8. A function f : X → R is S-measurable if and only if there exists a countable
subset D ⊆ X such that f is constant on X \ D.
Proof. We prove both directions:
(⇒) Measurable implies constant outside countable set
Assume f is S-measurable. Consider for each r ∈ R the set:
f −1 ({r}) = {x ∈ X : f (x) = r}
Since {r} is Borel and f is measurable, f −1 ({r}) ∈ S. There are two cases:
(1) Uncountable value set: Suppose the range f (X) is uncountable. Then:
• Only countably many r can have f −1 ({r}) uncountable (since their union must be X)
• For all other r, f −1 ({r}) must be countable
Let D be the union of all countable fibers f −1 ({r}). Then:
• D is countable (countable union of countable sets)
• On X \ D, f takes only values whose fibers are uncountable
• But there can be only one such value r0 because two uncountable sets in S must have
uncountable intersection
 13

Thus f ≡ r0 on X \ D.
(2) Countable value set: If f (X) is countable, let D = r∈f (X) f −1 ({r}). Then:
S

• D = X (so X \ D = ∅)
• The condition holds vacuously
(⇐) Constant outside countable set implies measurable
Assume f is constant on X \ D for some countable D, with f ≡ c on X \ D. For any Borel set
B ⊆ R:
• If c ∈ B, then f −1 (B) = DB ∪ (X \ D) where DB = {x ∈ D : f (x) ∈ B}
– DB is countable (subset of D)
– X \ D has countable complement (D is countable)
– Thus f −1 (B) ∈ S
/ B, then f −1 (B) = DB (countable) ∈ S
• If c ∈

Proof. We prove both directions of the equivalence.

(⇒) Suppose f : X → R is S-measurable. Then for every Borel set B ⊂ R, the preimage
f −1 (B) ∈ S.
Assume, for contradiction, that f takes more than one value on an uncountable subset of X.
That is, there exist distinct a, b ∈ R such that both f −1 ({a}) and f −1 ({b}) are uncountable. But
then neither of these sets is countable, and their complements are also uncountable, so they are
not in S. This contradicts the assumption that f is S-measurable.
Hence, f takes at most one value, say c ∈ R, on the co-countable set X \ D for some countable
D ⊂ X. Therefore, f (x) = c for all x ∈ X \ D.

(⇐) Conversely, suppose there exists a countable subset D ⊂ X such that f (x) = c for all
x ∈ X \ D. Let B ⊂ R be a Borel set. Consider the preimage f −1 (B).
/ B, then f −1 (B) ⊆ D, which is countable, so f −1 (B) ∈ S.
If c ∈
If c ∈ B, then
f −1 (B) = (X \ D) ∪ (f −1 (B) ∩ D),
where X \ D is co-countable and f −1 (B) ∩ D is countable. Hence, f −1 (B) is the union of a
co-countable and a countable set, and thus belongs to S.
In either case, f −1 (B) ∈ S, so f is S-measurable.

(1) Suppose B ⊆ R and f : B → R is increasing. Prove that f is continuous at every element
of B except for a countable subset.
Proof. For an increasing function, discontinuities can only be jump discontinuities. At each
discontinuity point x ∈ B, there exist:
lim f (y) = f (x− ) and lim f (y) = f (x+ )
y→x− y→x+

with f (x− ) ≤ f (x) ≤ f (x+ ) and f (x− ) < f (x+ ).

For each discontinuity, choose a rational rx ∈ (f (x− ), f (x+ )). Since f is increasing:
• Different discontinuities correspond to disjoint rational intervals
• The rationals are countable
Thus, the set of discontinuities is countable. 
(2) Suppose f : R → R is strictly increasing. Prove f −1 : f (R) → R is continuous.
14

Proof. We show f −1 is continuous at each y0 = f (x0 ) ∈ f (R):

Given  > 0, let:
y1 = f (x0 − ), y2 = f (x0 + )
Since f is strictly increasing:
y1 < y0 < y2
Take δ = min(y0 − y1 , y2 − y0 ). Then for y ∈ f (R) with |y − y0 | < δ:
f −1 (y) ∈ (x0 − , x0 + )
Thus f −1 is continuous at y0 . 

(3) Suppose f : R → R is strictly increasing and B ⊆ R is Borel. Prove f (B) is Borel.

Proof. From part (b), f −1 is continuous. Since:

• Continuous functions are Borel measurable
• f = (f −1 )−1 as relations
• The image f (B) = (f −1 )−1 (B)
and the Borel σ-algebra is preserved under continuous preimages, f (B) is Borel.
Alternatively, note that strictly increasing functions map:
• Open intervals to open intervals
• Closed intervals to closed intervals
• Thus preserve the generating sets of the Borel algebra


(1) (a) Let B ⊆ R and let f : B → R be an increasing function. We claim that f is continuous
at all but countably many points in B.

Proof. For x ∈ B, define the left and right limits:

f (x− ) := lim− f (y), f (x+ ) := lim+ f (y),
y→x y→x

where limits are taken over points in B.

Since f is increasing, these limits exist (possibly infinite). If f is discontinuous at x,
then either f (x− ) < f (x) or f (x+ ) > f (x). Each such discontinuity corresponds to a
jump.
Associate to each discontinuity a rational number in the interval (f (x− ), f (x)) or
(f (x), f (x+ )). These intervals are disjoint for distinct points of discontinuity, so there
can only be countably many such intervals (since there are only countably many
rational numbers). Hence, the set of discontinuities is countable. 

(b) Suppose f : R → R is strictly increasing. Define the inverse f −1 : f (R) → R.

Proof. Let yn → y in f (R). Since f is strictly increasing, it is injective and has an

inverse.
We want to show that f −1 (yn ) → f −1 (y). Suppose not. Then there exists an ε > 0
and a subsequence ynk such that |f −1 (ynk ) − f −1 (y)| ≥ ε for all k.
But since f is strictly increasing, f −1 is also monotonic (increasing), and hence it must
be continuous. This contradiction shows that f −1 is continuous. 

(c) Suppose f : R → R is strictly increasing and B ⊆ R is Borel.

 15

Proof. Since f is strictly increasing, it is measurable and Borel-measurable, though

possibly not continuous.
Consider the inverse function f −1 : f (R) → R, which is continuous by part (b). Then,
since B is Borel in R and f is injective, we can write
f (B) = y ∈ f (R) : f −1 (y) ∈ B = f −1−1 (B).


Since f −1 is continuous and B is Borel, f −1−1 (B) is Borel in f (R). But the Borel
σ-algebra on f (R) coincides with the trace σ-algebra induced from R. Hence, f (B) is
Borel in R.