Section 2.

3 Differentiable Functions of One Variable 73

2.3 DIFFERENTIABLE FUNCTIONS OF ONE VARIABLE

In calculus you studied differentiation, emphasizing rules for calculating derivatives. Here
we consider the theoretical properties of differentiable functions. In doing this, we assume
that you know how to differentiate elementary functions such as x n , e x , and sin x, and we
will use such functions in examples.

Definition of the Derivative

Definition 2.3.1 A function f is differentiable at an interior point x0 of its domain if

the difference quotient
f .x/ f .x0 /
; x ¤ x0 ;
x x0
approaches a limit as x approaches x0 , in which case the limit is called the derivative of f
at x0, and is denoted by f 0 .x0 /; thus,
f .x/ f .x0 /
f 0 .x0 / D lim : (1)
x!x0 x x0
It is sometimes convenient to let x D x0 C h and write (1) as
f .x0 C h/ f .x0 /
f 0 .x0 / D lim :
h!0 h
If f is defined on an open set S , we say that f is differentiable on S if f is differentiable
at every point of S . If f is differentiable on S , then f 0 is a function on S . We say that
f is continuously differentiable on S if f 0 is continuous on S . If f is differentiable on a
neighborhood of x0 , it is reasonable to ask if f 0 is differentiable at x0 . If so, we denote the
derivative of f 0 at x0 by f 00 .x0 /. This is the second derivative of f at x0 , and it is also
denoted by f .2/ .x0 /. Continuing inductively, if f .n 1/ is defined on a neighborhood of
x0 , then the nth derivative of f at x0, denoted by f .n/ .x0 /, is the derivative of f .n 1/ at
x0 . For convenience we define the zeroth derivative of f to be f itself; thus
f .0/ D f:

We assume that you are familiar with the other standard notations for derivatives; for
example,
f .2/ D f 00 ; f .3/ D f 000 ;
74 Chapter 2 Differential Calculus of Functions of One Variable

and so on, and

d nf
D f .n/ :
dx n

Example 2.3.1 If n is a positive integer and

f .x/ D x n ;

then
n 1
f .x/ f .x0 / xn x0n x x0 X n k 1 k
D D x x0 ;
x x0 x x0 x x0
kD0
so
n
X1
f 0 .x0 / D lim xn k 1 k
x0 D nx0n 1
:
x!x0
kD0

Since this holds for every x0 , we drop the subscript and write

d n
f 0 .x/ D nx n 1
or .x / D nx n 1
:
dx

To derive differentiation formulas for elementary functions such as sin x, cos x, and e x
directly from Definition 2.3.1 requires estimates based on the properties of these functions.
Since this is done in calculus, we will not repeat it here.

Interpretations of the Derivative

If f .x/ is the position of a particle at time x ¤ x0 , the difference quotient
f .x/ f .x0 /
x x0
is the average velocity of the particle between times x0 and x. As x approaches x0 , the
average applies to shorter and shorter intervals. Therefore, it makes sense to regard the
limit (1), if it exists, as the particle’s instantaneous velocity at time x0 . This interpretation
may be useful even if x is not time, so we often regard f 0 .x0 / as the instantaneous rate of
change of f .x/ at x0 , regardless of the specific nature of the variable x. The derivative also
has a geometric interpretation. The equation of the line through two points .x0 ; f .x0 // and
.x1 ; f .x1 // on the curve y D f .x/ (Figure 2.3.1) is
f .x1 / f .x0 /
y D f .x0 / C .x x0 /:
x1 x0
Varying x1 generates lines through .x0 ; f .x0// that rotate into the line

y D f .x0 / C f 0 .x0 /.x x0 / (2)

 Section 2.3 Differentiable Functions of One Variable 75

as x1 approaches x0. This is the tangent to the curve y D f .x/ at the point .x0 ; f .x0 //.
Figure 2.3.2 depicts the situation for various values of x1 .
y

y = f (x)

x
x0 x1

Figure 2.3.1

y = f (x)
e
lin
nt
ge
n
Ta

x
x0 x''1 x1 x1

Figure 2.3.2

Here is a less intuitive definition of the tangent line: If the function

T .x/ D f .x0 / C m.x x0 /

approximates f so well near x0 that

f .x/ T .x/
lim D 0;
x!x0 x x0
we say that the line y D T .x/ is tangent to the curve y D f .x/ at .x0 ; f .x0//.
 76 Chapter 2 Differential Calculus of Functions of One Variable

This tangent line exists if and only if f 0 .x0 / exists, in which case m is uniquely determined
by m D f 0 .x0 / (Exercise 1). Thus, (2) is the equation of the tangent line.
We will use the following lemma to study differentiable functions.

Lemma 2.3.2 If f is differentiable at x0; then

f .x/ D f .x0 / C Œf 0 .x0 / C E.x/.x x0 /; (3)

where E is defined on a neighborhood of x0 and

lim E.x/ D E.x0 / D 0:

x!x0

Proof Define
8
< f .x/ f .x0 /
f 0 .x0 /; x 2 Df and x ¤ x0 ;
E.x/ D x x0 (4)
:
0; x D x0 :

Solving (4) for f .x/ yields (3) if x ¤ x0 , and (3) is obvious if x D x0 . Definition 2.3.1
implies that limx!x0 E.x/ D 0. We defined E.x0 / D 0 to make E continuous at x0 .
Since the right side of (3) is continuous at x0 , so is the left. This yields the following
theorem.

Theorem 2.3.3 If f is differentiable at x0 ; then f is continuous at x0 :

The converse of this theorem is false, since a function may be continuous at a point
without being differentiable at the point.

Example 2.3.2 The function

f .x/ D jxj
can be written as

f .x/ D x; x > 0; (5)

or as
f .x/ D x; x < 0: (6)

From (5),

f 0 .x/ D x; x > 0;
and from (6),
f 0 .x/ D x; x < 0:

Neither (5) nor (6) holds throughout any neighborhood of 0, so neither can be used alone
to calculate f 0 .0/. In fact, since the one-sided limits

f .x/ f .0/ x
lim D lim (7)
x!0C x 0 x!0C x

and
f .x/ f .0/ x
lim D lim D 1 (8)
x!0 x 0 x!0 x
 Section 2.3 Differentiable Functions of One Variable 77

are different,
f .x/ f .0/
lim
x!0 x 0
does not exist (Theorem 2.1.6); thus, f is not differentiable at 0, even though it is continu-
ous at 0.

Interchanging Differentiation and Arithmetic Operations

The following theorem should be familiar from calculus.

Theorem 2.3.4 If f and g are differentiable at x0 ; then so are f C g; f g; and fg;

with
(a) .f C g/0 .x0 / D f 0.x0 / C g0 .x0 /I
(b) .f g/0 .x0 / D f 0.x0 / g.x0 /I
(c) .fg/0 .x0 / D f 0.x0 /g.x0 / C f .x0 /g0 .x0 /:
The quotient f =g is differentiable at x0 if g.x0 / ¤ 0; with
 0
f f 0 .x0 /g.x0 / f .x0 /g0 .x0 /
(d) .x0 / D :
g Œg.x0 /2

Proof The proof is accomplished by forming the appropriate difference quotients and
applying Definition 2.3.1 and Theorem 2.1.4. We will prove (c) and leave the rest to you
(Exercises 9, 10, and 11).
The trick is to add and subtract the right quantity in the numerator of the difference
quotient for .fg/0 .x0 /; thus,
f .x/g.x/ f .x0 /g.x0 / f .x/g.x/ f .x0 /g.x/ C f .x0 /g.x/ f .x0 /g.x0 /
D
x x0 x x0
f .x/ f .x0 / g.x/ g.x0 /
D g.x/ C f .x0 / :
x x0 x x0
The difference quotients on the right approach f 0 .x0 / and g0 .x0 / as x approaches x0 , and
limx!x0 g.x/ D g.x0 / (Theorem 2.3.3). This proves (c).

The Chain Rule

Here is the rule for differentiating a composite function.

Theorem 2.3.5 (The Chain Rule) Suppose that g is differentiable at x0 and f

is differentiable at g.x0 /: Then the composite function h D f ı g; defined by
h.x/ D f .g.x//;
is differentiable at x0; with
h0 .x0 / D f 0 .g.x0 //g0 .x0 /:
78 Chapter 2 Differential Calculus of Functions of One Variable

Proof Since f is differentiable at g.x0 /, Lemma 2.3.2 implies that

f .t/ f .g.x0 // D Œf 0 .g.x0 // C E.t/Œt g.x0 /;

where
lim E.t/ D E.g.x0 // D 0: (9)
t !g.x0 /

Letting t D g.x/ yields

f .g.x// f .g.x0 // D Œf 0 .g.x0 // C E.g.x//Œg.x/ g.x0 /:

Since h.x/ D f .g.x//, this implies that

h.x/ h.x0 / g.x/ g.x0 /
D Œf 0 .g.x0 / C E.g.x// : (10)
x x0 x x0
Since g is continuous at x0 (Theorem 2.3.3), (9) and Theorem 2.2.7 imply that

lim E.g.x// D E.g.x0 // D 0:

x!x0

Therefore, (10) implies that

h.x/ h.x0 /
h0 .x0 / D lim D f 0 .g.x0 //g0 .x0 /;
x!x0 x x0
as stated.

Example 2.3.3 If
1
f .x/ D sin x and g.x/ D ; x ¤ 0;
x
then
1
h.x/ D f .g.x// D sin ; x ¤ 0;
x
and   
0 0 1 1
h .x/ D f .g.x//g.x/ D cos ; x ¤ 0:
x x2

It may seem reasonable to justify the chain rule by writing

h.x/ h.x0 / f .g.x// f .g.x0 //
D
x x0 x x0
f .g.x// f .g.x0 // g.x/ g.x0 /
D
g.x/ g.x0 / x x0
and arguing that
f .g.x// f .g.x0 //
lim D f 0 .g.x0 //
x!x0 g.x/ g.x0 /
 Section 2.3 Differentiable Functions of One Variable 79

(because limx!x0 g.x/ D g.x0 // and

g.x/ g.x0 /
lim D g0 .x0 /:
x!x0 x x0
However, this is not a valid proof (Exercise 13).

One-Sided Derivatives
One-sided limits of difference quotients such as (7) and (8) in Example 2.3.2 are called one-
sided or right- and left-hand derivatives. That is, if f is defined on Œx0 ; b/, the right-hand
derivative of f at x0 is defined to be
f .x/ f .x0 /
fC0 .x0 / D lim
x!x0 C x x0
if the limit exists, while if f is defined on .a; x0, the left-hand derivative of f at x0 is
defined to be
f .x/ f .x0 /
f 0 .x0 / D lim
x!x0 x x0
if the limit exists. Theorem 2.1.6 implies that f is differentiable at x0 if and only if fC0 .x0 /
and f 0 .x0 / exist and are equal, in which case
f 0 .x0 / D fC0 .x0 / D f 0 .x0 /:

In Example 2.3.2, fC0 .0/ D 1 and f 0 .0/ D 1.

Example 2.3.4 If
8 3
<x ; x  0;
f .x/ D (11)
:x 2 sin 1 ; x > 0;
x
then 8 2
<3x ; x < 0;
f 0 .x/ D (12)
:2x sin 1 cos 1 ; x > 0:
x x
Since neither formula in (11) holds for all x in any neighborhood of 0, we cannot simply
differentiate either to obtain f 0 .0/; instead, we calculate
x 2 sin 1x 0 1
fC0 .0/ D lim D lim x sin D 0;
x!0C x 0 x!0C x
x3 0
f 0 .0/ D lim D lim x 2 D 0I
x!0 x 0 x!0
0
hence, f .0/ D fC0 .0/ D f 0 .0/ D 0.
80 Chapter 2 Differential Calculus of Functions of One Variable

This example shows that there is a difference between a one-sided derivative and a one-
sided limit of a derivative, since fC0 .0/ D 0, but, from (12), f 0 .0C/ D limx!0C f 0 .x/
does not exist. It also shows that a derivative may exist in a neighborhood of a point x0
(D 0 in this case), but be discontinuous at x0 .
Exercise 4 justifies the method used in
Example 2.3.4 to compute f 0 .x/ for x ¤ 0.

Definition 2.3.6
(a) We say that f is differentiable on the closed interval Œa; b if f is differentiable on
the open interval .a; b/ and fC0 .a/ and f 0 .b/ both exist.
(b) We say that f is continuously differentiable on Œa; b if f is differentiable on Œa; b,
f 0 is continuous on .a; b/, fC0 .a/ D f 0 .aC/, and f 0 .b/ D f 0 .b /.

Extreme Values
We say that f .x0 / is a local extreme value of f if there is a ı > 0 such that f .x/ f .x0 /
does not change sign on
.x0 ı; x0 C ı/ \ Df : (13)
More specifically, f .x0 / is a local maximum value of f if

f .x/  f .x0 / (14)

or a local minimum value of f if

f .x/  f .x0 / (15)

for all x in the set (13). The point x0 is called a local extreme point of f , or, more specifi-
cally, a local maximum or local minimum point of f .

x
−1 −1 1 1 2 3 4
2 2

Figure 2.3.3
 Section 2.3 Differentiable Functions of One Variable 81

Example 2.3.5 If
8̂ 1
ˆ 1; 1<x 2
ˆ
ˆ
< 1
jxj; < x  12 ;
f .x/ D 2
ˆ
ˆ
ˆ 1 x 1
:̂ p sin ; 2
<x4
2 2

(Figure 2.3.3), then 0, 3, and every x in . 1; 12 / are local minimum points of f , while 1,
4, and every x in . 1; 21  are local maximum points.

It is geometrically plausible that if the curve y D f .x/ has a tangent at a local extreme
point of f , then the tangent must be horizontal; that is, have zero slope. (For example, in
Figure 2.3.3, see x D 1, x D 3, and every x in . 1; 1=2/.) The following theorem shows
that this must be so.

Theorem 2.3.7 If f is differentiable at a local extreme point x0 2 Df0 ; then f 0 .x0 / D 0:

If f 0 .x0 / D 0, we say that x0 is a critical point of f . Theorem 2.3.7 says that every
local extreme point of f at which f is differentiable is a critical point of f . The converse
is false. For example, 0 is a critical point of f .x/ D x 3 , but not a local extreme point.

Rolle’s Theorem
The use of Theorem 2.3.7 for finding local extreme points is covered in calculus, so we will
not pursue it here. However, we will use Theorem 2.3.7 to prove the following fundamental
theorem, which says that if a curve y D f .x/ intersects a horizontal line at x D a and
x D b and has a tangent at .x; f .x// for every x in .a; b/, then there is a point c in .a; b/
such that the tangent to the curve at .c; f .c// is horizontal (Figure 2.3.4, page 82).
82 Chapter 2 Differential Calculus of Functions of One Variable
y

x
a c b

Figure 2.3.4

Theorem 2.3.8 (Rolle’s Theorem) Suppose that f is continuous on the closed

interval Œa; b and differentiable on the open interval .a; b/; and f .a/ D f .b/: Then
f 0 .c/ D 0 for some c in the open interval .a; b/:

Proof Since f is continuous on Œa; b, f attains a maximum and a minimum value on
Œa; b (Theorem 2.2.9). If these two extreme values are the same, then f is constant on
.a; b/, so f 0 .x/ D 0 for all x in .a; b/. If the extreme values differ, then at least one must
be attained at some point c in the open interval .a; b/, and f 0 .c/ D 0, by Theorem 2.3.7.

Intermediate Values of Derivatives

A derivative may exist on an interval Œa; b without being continuous on Œa; b. Neverthe-
less, an intermediate value theorem similar to Theorem 2.2.10 applies to derivatives.

Theorem 2.3.9 (Intermediate Value Theorem for Derivatives) Suppose

that f is differentiable on Œa; b; f 0 .a/ ¤ f 0 .b/; and  is between f 0 .a/ and f 0 .b/: Then
f 0 .c/ D  for some c in .a; b/:
 Section 2.3 Differentiable Functions of One Variable 83

Mean Value Theorems

Theorem 2.3.10 (Generalized Mean Value Theorem) If f and g are con-

tinuous on the closed interval Œa; b and differentiable on the open interval .a; b/; then
Œg.b/ g.a/f 0 .c/ D Œf .b/ f .a/g0 .c/ (20)
for some c in .a; b/:
Proof The function
h.x/ D Œg.b/ g.a/f .x/ Œf .b/ f .a/g.x/
is continuous on Œa; b and differentiable on .a; b/, and
h.a/ D h.b/ D g.b/f .a/ f .b/g.a/:
Therefore, Rolle’s theorem implies that h0 .c/ D 0 for some c in .a; b/. Since
h0 .c/ D Œg.b/ g.a/f 0 .c/ Œf .b/ f .a/g0 .c/;
this implies (20).
The following special case of Theorem 2.3.10 is important enough to be stated separately.

Theorem 2.3.11 (Mean Value Theorem) If f is continuous on the closed

interval Œa; b and differentiable on the open interval .a; b/; then
f .b/ f .a/
f 0 .c/ D
b a
for some c in .a; b/:

Proof Apply Theorem 2.3.10 with g.x/ D x.

Theorem 2.3.11 implies that the tangent to the curve y D f .x/ at .c; f .c// is parallel to
the line connecting the points .a; f .a// and .b; f .b// on the curve (Figure 2.3.5, page 84).

Consequences of the Mean Value Theorem

If f is differentiable on .a; b/ and x1 , x2 2 .a; b/ then f is continuous on the closed
interval with endpoints x1 and x2 and differentiable on its interior. Hence, the mean value
theorem implies that
f .x2 / f .x1 / D f 0 .c/.x2 x1/
for some c between x1 and x2 . (This is true whether x1 < x2 or x2 < x1.) The next three
theorems follow from this.
84 Chapter 2 Differential Calculus of Functions of One Variable

Theorem 2.3.12 If f 0.x/ D 0 for all x in .a; b/; then f is constant on .a; b/:

Theorem 2.3.13 If f 0 exists and does not change sign on .a; b/; then f is monotonic
on .a; b/ W increasing; nondecreasing; decreasing; or nonincreasing as
f 0 .x/ > 0; f 0 .x/  0; f 0 .x/ < 0; or f 0 .x/  0;
respectively; for all x in .a; b/:

Theorem 2.3.14 If
jf 0 .x/j  M; a < x < b;
then
jf .x/ f .x 0 /j  M jx x 0 j; x; x 0 2 .a; b/: (21)

A function that satisfies an inequality like (21) for all x and x 0 in an interval is said to
satisfy a Lipschitz condition on the interval.

f (b) y = f (x)

f (c)

f (a)

x
a c b

Figure 2.3.5