Chapter 4

Functional Limits and

Continuity

4.1. Functional Limits

Definition 4.1 (-δ definition of functional limits). Let f : A → R, and let c be a limit
point of the domain A. We say the limit of f as x approaches c is a number L and write
limx→c f (x) = L provided that, for each  > 0, there exists a δ > 0 such that whenever
0 < |x − c| < δ and x ∈ A it follows that |f (x) − L| < .

Note that the condition 0 < |x − c| is simply saying x 6= c. Therefore, as known from
Calculus, the limit value L has nothing to do with whether f is defined at c or not; even
f (c) is defined (meaning c ∈ A), L may not have any relation with it.
Example 4.1. (i) Let f (x) = 3x + 1. In this case the domain A of this formula-defined
function is considered to be all real numbers, that certainly makes sense of f (x). Show
limx→2 f (x) = 7.

Proof. Let  > 0. We need to produce a δ > 0 with the property that |f (x) − 7| <  holds
for all x satisfying 0 < |x − 2| < δ. The ending requirement is the inequality |f (x) − 7| < ,
which can be rewritten as
|f (x) − 7| = |(3x + 1) − 7| = |3x − 6| = 3|x − 2| < .
Hence the requirement |f (x) − 7| <  is equivalent to |x − 2| < /3; that is, whenever
|x − 2| < /3, it follows that |f (x) − 7| < . Therefore, we can select simply δ = /3 > 0 to
satisfy the definition. 

(ii) Show limx→2 x2 = 4.

Proof. Given  > 0, our goal is to produce a δ > 0 such that |x2 − 4| <  for all x with
0 < |x−2| < δ. As above the domain set A = R. We start to analyze the ending requirement
|x2 − 4| < , which can be rewritten as
|x2 − 4| = |x + 2||x − 2| < .
Unlike the previous example, in front of |x − 2| there is a function |x + 2|, not simply a
constant number; we cannot divide by |x + 2| to simply select δ = /|x + 2| since this δ

1
2 4. Functional Limits and Continuity

depends on x. The idea is to first choose one fixed δ to control the term |x+2|. For example,
let |x − 2| < 1 (with δ1 = 1). For all such x’s, we have −1 < x − 2 < 1 and hence 1 < x < 3
and so 3 < x + 2 < 5; that is, |x + 2| < 5. Hence, if |x − 2| < 1 then |x + 2| < 5 and so
|x2 − 4| = |x + 2||x − 2| ≤ 5|x − 2| (note that this inequality holds when |x − 2| < 1).
Therefore, for all such x’s, to make |x2 − 4| < , it suffices to require |x − 2| < /5 = δ2 .
Now, choose δ = min{1, /5}. If 0 < |x − 2| < δ, then both |x − 2| < δ1 = 1 and
|x − 2| < δ2 = /5 hold and hence

|x2 − 4| = |x + 2||x − 2| ≤ 5|x − 2| < 5 × = ,
5
and the limit is proved. 

Topological Version of Functional Limits. Since the statement |f (x) − L| <  is equiv-
alent to f (x) ∈ V (L) and the statement |x − c| < δ is equivalent to x ∈ Vδ (c) and hence the
statement 0 < |x − c| < δ and x ∈ A simply means x ∈ (Vδ (c) \ {c}) ∩ A, we can rephrase
the -δ definition above by using the topological terminologies (of neighborhoods).
Definition 4.2 (Topological Definition of Functional Limits). Let f : A → R, and
let c be a limit point of the domain A. We say limx→c f (x) = L provided that
∀  > 0 ∃ δ > 0, f (V̂δ (c) ∩ A) ⊆ V (L),
where V̂δ (c) = Vδ (c) \ {c} denotes the punctured neighborhood of c.

Sequential Criterion for Functional Limits. Functional limits can be completely char-
acterized by the convergence of all related sequences.
Theorem 4.1 (Sequential Criterion for Functional Limits). Let f : A → R and c be
a limit point of A. Then the following two conditions are equivalent:
(i) lim f (x) = L.
x→c
(ii) For all sequences (xn ) satisfying xn ∈ A, xn 6= c and (xn ) → c, it follows that the
sequence (f (xn )) → L.

Proof. 1. First assume (i) and we prove (ii). Let (xn ) satisfy xn ∈ A, xn 6= c and (xn ) → c.
Given each  > 0, ∃ δ > 0 such that |f (x) − L| < δ for all x ∈ A with 0 < |x − c| < δ. For
this δ > 0, ∃ N ∈ N such that |xn − c| < δ for all n ≥ N in N. Since xn 6= c, we have
0 < |xn − c| < δ for all n ≥ N and thus |f (xn ) − L| < . This proves (f (xn )) → L.
2. We now assume (ii) and prove (i). Suppose lim f (x) 6= L. Then ∃ 0 > 0, xn ∈ A
x→c
with 0 < |xn − c| < 1/n but |f (xn ) − L| ≥ 0 for all n = 1, 2, · · · . (This is the negation of
lim f (x) = L.) For this sequence (xn ), by (ii), (f (xn )) → L, a contradiction to |f (xn )−L| ≥
x→c
0 for all n = 1, 2, · · · . 
Corollary 4.2 (Divergence Criterion for Functional Limits). Let f : A → R and c be a
limit point of A. If there exist two sequences (xn ) and (yn ) in A, with xn 6= c and yn 6= c,
satisfying the property
lim xn = lim yn = c but lim f (xn ) 6= lim f (yn ),
then the functional limit limx→c f (x) does not exist.
Example 4.2. Show that limx→0 sin(1/x) does not exist.
4.2. Combinations of Continuous Functions 3

Proof. Let xn = 1/2nπ and yn = 1/(2nπ + π/2) for all n ∈ N. Then xn 6= 0 and yn 6= 0
and lim xn = lim yn = 0. But sin(1/xn ) = sin(2nπ) = 0 while sin(1/yn ) = sin(2nπ + π/2) =
sin(π/2) = 1. By the corollary above, the functional limit limx→0 sin(1/x) does not exist. 
Theorem 4.3 (Algebraic Limit Theorem for Functional Limits). Let f, g : A → R
and c be a limit point of A. Assume lim f (x) = L and lim g(x) = M exist. Then
x→c x→c
(i) lim [af (x) + bg(x)] = aL + bM for all a, b ∈ R,
x→c
(ii) lim [f (x)g(x)] = LM,
x→c
(iii) lim [f (x)/g(x)] = L/M, provided M 6= 0.
x→c

4.2. Combinations of Continuous Functions

Definition 4.3 (Continuous Functions). Let f : A → R and c ∈ A.
(i) We say f is continuous at a point c ∈ A if, for each  > 0, there exists a δ > 0
such that whenever |x − c| < δ and x ∈ A it follows that |f (x) − f (c)| < .
If f is not continuous at c we say f is discontinuous at c.
(ii) We say f is a continuous function on A if it is continuous at every point in A.

Note that continuity at c is not defined if f (c) is not defined, i.e., if c ∈

/ A. If c ∈ A is
an isolated point of A, then f is always continuous at c since for some δ > 0 the only point
x satisfying |x − c| < δ and x ∈ A is x = c and hence the condition |f (x) − f (c)| = 0 < 
always holds. If c ∈ A is a limit point of A, then continuity of f at c is simply equivalent to
lim f (x) = f (c).
x→c
This is the most interesting case.
Theorem 4.4 (Characterizations of Continuity). Let f : A → R and c ∈ A be a limit
point of A. Then the following conditions are equivalent:
(i) f is continuous at c.
(ii) lim f (x) = f (c).
x→c
(iii) ∀  > 0 ∃ δ > 0, f (Vδ (c) ∩ A) ⊆ V (f (c)).
(iv) Whenever xn ∈ A and (xn ) → c it follows that (f (xn )) → f (c).

Proof. (i), (ii), and (iii) are simply a different way to describe the definition of the con-
tinuity; the condition (iv) with xn 6= c would be already equivalent to the convergence
limx→c f (x) = f (c). Details are omitted. 
Corollary 4.5 (Criterion for Discontinuity). Let f : A → R and c ∈ A be a limit point
of A. Then f is not continuous at c if and only if for some number 0 > 0 and sequence
(xn ) in A with (xn ) → c it follows that |f (xn ) − f (c)| ≥ 0 for all n ∈ N.

Proof. Use ¬(ii) ⇐⇒ ¬(iv). 

Theorem 4.6 (Algebraic Continuity Theorem). Let f, g : A → R be continuous at a
point c ∈ A. Then
(i) af (x) + bg(x) is continuous at c for all a, b ∈ R;
(ii) f (x)g(x) is continuous at c;
(iii) f (x)/g(x) is continuous at c, provided the quotient is well defined.
4 4. Functional Limits and Continuity

Example 4.3. Polynomials are all continuous functions on R. Hence all rational functions
(quotients of polynomials) are continuous at points where the denominator is not zero.
Example 4.4. Let (
x sin(1/x) if x 6= 0
g(x) =
0 if x = 0.
Then g is continuous at 0.

Proof. Note that |g(x)| ≤ |x| for all x ∈ R; this is clearly true if x = 0 and is also true if
x 6= 0 by the formula of g(x) since | sin θ| ≤ 1 for all θ. Given each  > 0, let δ = . Then,
whenever |x − 0| < δ, since g(0) = 0,
|g(x) − g(0)| = |g(x)| ≤ |x| < δ = .
So g is continuous at 0. 
√
Example 4.5. Function f (x) = x is continuous on A = {x ∈ R : x ≥ 0} = [0, ∞). This
can be shown by the definition directly or using the sequential criterion.
Theorem 4.7 (Composition of Continuous Functions). Let f : A → R and g : B → R,
where B ⊇ f (A) and so that the composition function g ◦ f : A → R is defined by
g ◦ f (x) = g(f (x)) for x ∈ A. Then, if f is continuous at a point c ∈ A and g is continuous
at f (c) ∈ B, g ◦ f is continuous at c ∈ A as well.

Proof. Given  > 0, there exists a τ > 0 such that

|g(y) − g(f (c))| <  ∀ y ∈ B, |y − f (c)| < τ.
With this τ > 0, there exists a δ > 0 such that
|f (x) − f (c)| < τ ∀ x ∈ A, |x − c| < δ.
Therefore, whenever x ∈ A and |x − c| < δ, it follows that f (x) ∈ B and |f (x) − f (c)| < τ ,
and hence
|g ◦ f (x) − g ◦ f (c)| = |g(f (x)) − g(f (c))| < .
Hence, by definition, g ◦ f is continuous at c. 

4.3. Continuous Functions on Compact Sets and Uniform Continuity

Definition 4.4. A function f : A → R is said to be bounded on a set B ⊆ A if the set
f (B) is a bounded set. If f (A) is a bounded set, we say f : A → R is a bounded function.
Theorem 4.8 (Preservation of Compact Sets). Let f : A → R be continuous on A. If
K ⊆ A is a compact set, then f (K) is compact as well.

Proof. Given any sequence (yn ) in f (K), we show that there exists a subsequence (ynk )
converging to some limit y ∈ f (K). This will prove that f (K) is compact. Since yn ∈ f (K),
we have yn = f (xn ) for some xn ∈ K. Now (xn ) is sequence contained in K. Since K is
compact, there exists a subsequence (xnk ) converging to a limit x ∈ K. By the continuity,
(f (xnk )) → f (x). Since ynk = f (xnk ), we have (ynk ) → y = f (x) ∈ f (K). This proves that
f (K) is compact. 
4.3. Continuous Functions on Compact Sets and Uniform Continuity 5

Remark 4.1. In contrast to this theorem, continuous functions do not preserve bounded
sets, open sets or closed sets. Consider the following examples.
(i) f (x) = x1 is continuous in (0, 1), but f is not bounded in (0, 1). This function also
maps the closed interval A = [1, ∞) to the set f (A) = (0, 1], which is not closed.
(ii) f (x) = x2 maps the open set A = (−1, 1) to the set f (A) = [0, 1), which is not open.
Theorem 4.9 (Extreme Value Theorem). If f : K → R is continuous on a compact
set K ⊆ R, then f attains its maximum and minimum values; namely, there exist numbers
x∗ , x∗ ∈ K such that f (x∗ ) ≤ f (x) ≤ f (x∗ ) for all x ∈ K.

Proof. We only prove the maximum case. Let A = f (K). Then A is compact; hence A
is bounded and closed. Let M = sup A. Then f (x) ≤ M for all x ∈ K. We now show
that there exists a number x∗ ∈ K such that f (x∗ ) = M, which proves that the maximum
is attained. Since M = sup A, for each n ∈ N, there exists a number yn ∈ A such that
yn > M − n1 (which just states that M − n1 is not an upper-bound for A). Since yn ≤ M ,
this implies (yn ) → M. Now, since yn ∈ A = f (K) and A is compact, by definition of
compact sets, (yn ) has a convergent subsequence whose limit is in A, but this subsequence
also converges to M ; therefore, M ∈ A = f (K), which means that there exists a number
x∗ ∈ K such that M = f (x∗ ). 

Uniform Continuity.
Definition 4.5. A function f : A → R is said to be uniformly continuous on A if, for
every  > 0, there exists a δ > 0 such that whenever x, y ∈ A and |x − y| < δ it follows that
|f (x) − f (y)| < .
Remark 4.2. We only talk about the uniform continuity of a function on a given set not
at a point. From the definition, we see that every uniformly continuous function on a set A
must be continuous at every point of A and so must be a continuous function on A.
Example 4.6. Show that f (x) = 3x + 1 is uniformly continuous on R.

Proof. Since f (x) − f (y) = 3(x − y), given each  > 0, letting δ = /3, it follows that
whenever |x − y| < δ,
|f (x) − f (y)| = 3|x − y| < 3δ = .
So f (x) = 3x + 1 is uniformly continuous on R. 
Example 4.7. Is function g(x) = x2 uniformly continuous on R?

Solution. Let us suppose that g(x) = x2 is uniformly continuous on R. Then ∀  > 0 ∃ δ > 0
such that |g(x) − g(y)| <  for all x, y ∈ R with |x − y| < δ. However, if we choose (large
numbers) x = N + δ/2 and y = N with N ∈ N and N > 2/δ, then
δ δ
|g(x) − g(y)| = |x − y||x + y| = (2N + ) > δN > 2,
2 2
a contradiction; therefore, g(x) = x2 is not uniformly continuous on R. 

By negating the definition of uniform continuity, we have the following criterion for
nonuniform continuity.
Theorem 4.10 (Sequential Criterion for Nonuniform Continuity). A function f : A →
R fails to be uniformly continuous on A if and only if there exist 0 > 0 and two sequences
(xn ) and (yn ) in A satisfying |xn − yn | → 0 but |f (xn ) − f (yn )| ≥ 0 .
6 4. Functional Limits and Continuity

Example 4.8. The function h(x) = sin(1/x) is continuous at every point in the open
interval (0, 1). Let
1 1
xn = , yn = in (0, 1);
2nπ 2nπ + π2
then xn − yn → 0 but |h(xn ) − h(yn )| = 1. So h(x) is not uniformly continuous on (0, 1).
Theorem 4.11 (Uniform continuity on compact sets). Every continuous function on
a compact set K is uniformly continuous on K.

Proof. Suppose f is not uniformly continuous on K. Then ∃ 0 > 0 ∃ xn , yn ∈ K such that

|xn − yn | → 0 but |f (xn ) − f (yn )| ≥ 0 . Since (xn ) is in K and K is compact, by definition
of compact sets, there exists a convergent subsequence (xnk ) converging to a number x ∈ K
as nk → ∞. We also have (ynk ) → x from |ynk − x| ≤ |ynk − xnk | + |xnk − x| → 0 as nk → ∞.
Hence, by continuity, f (xnk ) → f (x) and f (ynk ) → f (x) and so f (xnk ) − f (ynk ) → 0; this
contradicts with |f (xnk ) − f (ynk )| ≥ 0 . 
Remark 4.3. By Theorem 4.11, the function g(x) = x2 is uniformly continuous on all closed
intervals [a, b] but is not uniformly continuous on R (see Example 4.7). So, a function may
be uniformly continuous on one set but not uniformly continuous on another set.
Example 4.9. Let f be continuous on [0, ∞) and uniformly continuous on [b, ∞) for some
b > 0. Show that f is uniformly continuous on [0, ∞).

Proof. Use the uniform continuity of f on both [0, b] and [b, ∞). Details are left as Home-
work! 

4.4. The Intermediate Value Theorem

Theorem 4.12 (Intermediate Value Theorem (IVT)). Let f : [a, b] → R be continuous
and f (a) 6= f (b). Assume L is a number between f (a) and f (b). Then there exists a number
c ∈ (a, b) where f (c) = L.

Proof. We discuss the proof based on the AoC; other different proofs are also given in the
text. To fix the idea, we assume f (a) < L < f (b) (the other case f (b) < L < f (a) can be
treated similarly). Consider the set
K = {x ∈ [a, b] : f (x) ≤ L}.
Since a ∈ K, K is nonempty; since K ⊆ [a, b], K is bounded. Hence by the AoC, let
c = sup K. Since b is an upper-bound for K, we have c ≤ b; since a ∈ K, we have a ≤ c.
Hence c ∈ [a, b]. We prove that f (c) = L. Suppose f (c) 6= L; then the following two cases
would both reach a contradiction.
Case 1: f (c) < L. In this case, c 6= b and hence c < b. Take  = L − f (c) and V (f (c)) =
(f (c) − , f (c) + ). By continuity at c, there exists a δ > 0 such that f (Vδ (c) ∩ [a, b]) ⊆
V (f (c)). Take a number δ0 > 0 such that δ0 < min{b − c, δ}. Then c + δ0 ∈ [a, b] ∩ Vδ (c) and
so f (c + δ0 ) ∈ V (f (c)); that is, f (c + δ0 ) < f (c) +  = L. Hence c + δ0 ∈ K, a contradiction
with c = sup K.
Case 2: f (c) > L. In this case, c 6= a and hence c > a. Take  = f (c) − L and
V (f (c)) = (f (c) − , f (c) + ). By continuity at c, there exists a δ > 0 such that f (Vδ (c) ∩
[a, b]) ⊆ V (f (c)). Take a number δ0 > 0 such that δ0 < min{c − a, δ}; so c − δ0 > a.
Hence [c − δ0 , c] ⊆ [a, b] ∩ Vδ (c) and so f (x) ∈ V (f (c)) for all x ∈ [c − δ0 , c]; that is,
4.5. Sets of Discontinuity for Monotone Functions 7

f (x) > f (c) −  = L and hence x ∈/ K if x ∈ [c − δ0 , c]. This implies for all x ∈ K it must
follow that x < c − δ0 and hence c − δ0 is an upper-bound for K, which, by c = sup K,
would imply that c ≤ c − δ0 , again a contradiction. 

4.5. Sets of Discontinuity for Monotone Functions

One-sided Limits. Let f : (a, b) → R and c ∈ (a, b). We say lim f (x) = L if, for each
x→c+
 > 0, there exists a δ > 0 such that
|f (x) − L| <  whenever c < x < min{c + δ, b}.
Similarly, we say lim f (x) = L if, for each  > 0, there exists a δ > 0 such that
x→c−
|f (x) − L| <  whenever max{a, c − δ} < x < c.
Theorem 4.13. Let f : (a, b) → R and c ∈ (a, b). Then lim f (x) exists if and only if both
x→c
lim f (x) and lim f (x) exist and are equal. In this case, all these limits are the same.
x→c+ x→c−

Type of Discontinuity. Let f : (a, b) → R and c ∈ (a, b). If f is discontinuous at c, then

we have the following three cases:
(a) (Removable discontinuity) lim f (x) exists but is not equal to f (c).
x→c
(b) (Jump discontinuity) lim f (x) and lim f (x) both exist but are not equal.
x→c+ x→c−
(c) (Essential discontinuity) None of case (a) or (b) holds.
Definition 4.6. A function f : (a, b) → R is said to be increasing on (a, b) (or decreasing
on (a, b)) if f (t) ≤ f (s) (or f (t) ≥ f (s)) for all a < t < s < b. A function f : (a, b) → R is
called monotone on (a, b) if it is either increasing on (a, b) or decreasing on (a, b).
Theorem 4.14. The set of discontinuity of a monotone function on (a, b) is at most count-
able.

Proof. Without loss of generality, assume f : (a, b) → R is increasing. Let

S = {c ∈ (a, b) : f is discontinuous at c}.
Assume S =6 ∅. We show that at every c ∈ S the function f has a jump discontinuity. First
show that both lim f (x) and lim f (x) exist and satisfy that
x→c− x→c+
lim f (x) ≤ f (c) ≤ lim f (x).
x→c− x→c+
(Homework!) Therefore, if lim f (x) exists, then lim f (x) = f (c); that is f is continuous
x→c x→c
at c. So f can not have a removable discontinuity. Hence every point c ∈ S is a jump
discontinuity of f . We now define a function h : S → Q as follows. Given c ∈ S, since
lim f (x) < lim f (x), there exists a rational number r between lim f (x) and lim f (x).
x→c− x→c+ x→c− x→c+
Take any such rational number r and define h(c) = r. This defines a function h : S → Q.
We show h is one-to-one. Let c, d ∈ S and c < d. We show that h(c) < h(d). This follows
from the inequalities:
h(c) < lim f (x) ≤ lim f (x) < h(d).
x→c+ x→d−
(Homework!) Since h : S → Q is one-to-one and Q is countable, it follows that S is at
most countable.