Banner A. The Calculus Lifesaver: All the Tools You Need to Excel at Calculus
Подождите немного. Документ загружается.
86 • Continuity and Differentiability
Again, velocity can be negative while speed must be nonnegative. If the car
has a negative average velocity over a certain time period, then it has ended
to the left of where it began. If instead the average velocity is 0 over the time
period, then the car has ended up exactly where it began. Notice that in this
case the car might have a high average speed even though its average velocity
is 0! In general, just like displacement, if the car is going in just one direction,
then the average speed is just the absolute value of the average velocity.
5.2.3 Instantaneous velocity
We now revisit our crucial question in terms of velocity: how do you measure
the velocity of the car at a given instant? The idea, as we saw above, is
to take the average velocity of the car over smaller and smaller time periods
beginning at the instant the photo was taken. Here’s how it works in symbols.
Let t be the instant of time we care about. For example, if a race started at
2 p.m., you might decide to work in seconds with 0 representing the starting
time; in that case, if the photo was taken at 2:03 p.m. then you’d want to take
t = 180. Anyway, suppose that u is a short time later than t. Let’s write v
t↔u
to mean the average velocity of the car during the time interval beginning at
time t and ending at time u. Now we just push u closer and closer to t. How
close? As close as we can! That’s where the limit comes in. In fact,
instantaneous velocity at time t = lim
u→t
+
v
t↔u
.
Why neglect what happens before time t, though? We can make the above
definition a little more general by allowing u to be before t; then we can
replace the right-hand limit by a two-sided limit:
instantaneous velocity at time t = lim
u→t
v
t↔u
.
Now we need a few more formulas. Let’s suppose we know exactly where on
the highway the car is at any instant of time. In particular, suppose that at
time t, the car is at position f (t). That is, let
f(t) = position of car at time t.
We can now calculate the average velocity v
t↔u
exactly:
v
t↔u
=
position at time u − position at time t
u − t
=
f(u) − f(t)
u − t
.
Notice that the denominator u − t is the length of time involved (provided
that u is after
∗
t). Anyway, now we just take a limit as u → t:
instantaneous velocity at time t = lim
u→t
f(u) − f(t)
u − t
.
Of course, you cannot just substitute u = t in the previous limit, because then
you get the indeterminate form 0/0. You really do need to use limits.
∗
If u is before t, then the denominator should be t − u, but then the numerator should
be f(t) − f(u), so it all works out!
Section 5.2.4: The graphical interpretation of velocity • 87
One more little variation. Let’s define h = u − t. Then since u is very
close to t, the difference h between the two times must be very small. Indeed,
as u → t, we can see that h → 0. If we make this substitution in the above
limit, then because u = t + h, we also have
instantaneous velocity at time t = lim
h→0
f(t + h) − f(t)
h
.
There’s no real difference between this formula and the previous one; it’s just
written a little differently.
Let’s look at a quick example. Suppose that the car starts at rest at the
PSfrag
replacements
(
a, b)
[
a, b]
(
a, b]
[
a, b)
(
a, ∞)
[
a, ∞)
(
−∞, b)
(
−∞, b]
(
−∞, ∞)
{
x : a < x < b}
{
x : a ≤ x ≤ b}
{
x : a < x ≤ b}
{
x : a ≤ x < b}
{
x : x ≥ a}
{
x : x > a}
{
x : x ≤ b}
{
x : x < b}
R
a
b
shado
w
0
1
4
−
2
3
−
3
g(
x) = x
2
f(
x) = x
3
g(
x) = x
2
f(
x) = x
3
mirror
(y = x)
f
−
1
(x) =
3
√
x
y = h
(x)
y = h
−
1
(x)
y =
(x − 1)
2
−
1
x
Same
height
−
x
Same
length,
opp
osite signs
y = −
2x
−
2
1
y =
1
2
x − 1
2
−
1
y =
2
x
y =
10
x
y =
2
−x
y =
log
2
(x)
4
3
units
mirror
(x-axis)
y = |
x|
y = |
log
2
(x)|
θ radians
θ units
30
◦
=
π
6
45
◦
=
π
4
60
◦
=
π
3
120
◦
=
2
π
3
135
◦
=
3
π
4
150
◦
=
5
π
6
90
◦
=
π
2
180
◦
= π
210
◦
=
7
π
6
225
◦
=
5
π
4
240
◦
=
4
π
3
270
◦
=
3
π
2
300
◦
=
5
π
3
315
◦
=
7
π
4
330
◦
=
11
π
6
0
◦
=
0 radians
θ
hypotenuse
opp
osite
adjacen
t
0
(≡ 2π)
π
2
π
3
π
2
I
I
I
I
II
IV
θ
(
x, y)
x
y
r
7
π
6
reference
angle
reference
angle =
π
6
sin
+
sin −
cos
+
cos −
tan
+
tan −
A
S
T
C
7
π
4
9
π
13
5
π
6
(this
angle is
5π
6
clo
ckwise)
1
2
1
2
3
4
5
6
0
−
1
−
2
−
3
−
4
−
5
−
6
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
π
2
y =
sin(x)
1
0
−
1
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
5
π
2
2
π
2
π
3
π
2
π
π
2
y =
sin(x)
y =
cos(x)
−
π
2
π
2
y =
tan(x), −
π
2
<
x <
π
2
0
−
π
2
π
2
y =
tan(x)
−
2π
−
3π
−
5
π
2
−
3
π
2
−
π
−
π
2
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
y =
sec(x)
y =
csc(x)
y =
cot(x)
y = f(
x)
−
1
1
2
y = g(
x)
3
y = h
(x)
4
5
−
2
f(
x) =
1
x
g(
x) =
1
x
2
etc.
0
1
π
1
2
π
1
3
π
1
4
π
1
5
π
1
6
π
1
7
π
g(
x) = sin
1
x
1
0
−
1
L
10
100
200
y =
π
2
y = −
π
2
y =
tan
−1
(x)
π
2
π
y =
sin(
x)
x
,
x > 3
0
1
−
1
a
L
f(x) = x sin (1/x)
(0 < x < 0.3)
h(x) = x
g(x) = −x
a
L
lim
x→a
+
f(x) = L
lim
x→a
+
f(x) = ∞
lim
x→a
+
f(x) = −∞
lim
x→a
+
f(x) DNE
lim
x→a
−
f(x) = L
lim
x→a
−
f(x) = ∞
lim
x→a
−
f(x) = −∞
lim
x→a
−
f(x) DNE
M
}
lim
x→a
−
f(x) = M
lim
x→a
f(x) = L
lim
x→a
f(x) DNE
lim
x→∞
f(x) = L
lim
x→∞
f(x) = ∞
lim
x→∞
f(x) = −∞
lim
x→∞
f(x) DNE
lim
x→−∞
f(x) = L
lim
x→−∞
f(x) = ∞
lim
x→−∞
f(x) = −∞
lim
x→−∞
f(x) DNE
lim
x →a
+
f(x) = ∞
lim
x →a
+
f(x) = −∞
lim
x →a
−
f(x) = ∞
lim
x →a
−
f(x) = −∞
lim
x →a
f(x) = ∞
lim
x →a
f(x) = −∞
lim
x →a
f(x) DNE
y = f (x)
a
y =
|x|
x
1
−1
y =
|x + 2|
x + 2
1
−1
−2
1
2
3
4
a
a
b
y = x sin
1
x
y = x
y = −x
a
b
c
d
C
a
b
c
d
−1
0
1
2
3
7 mile marker, then accelerates to the right beginning at time t = 0 hours. It
turns out that the car’s position at time t might be something like 15t
2
+ 7
(the number 15 here depends on the acceleration). Without worrying about
why this is true, let’s just let f(t) = 15t
2
+7 and see if we can find the velocity
of the car at any time t.
Using the above formula, we have
instantaneous velocity at time t = lim
h→0
f(t + h) − f(t)
h
= lim
h→0
(15(t + h)
2
+ 7) − (15t
2
+ 7)
h
.
Now expand (t + h)
2
= t
2
+ 2th + h
2
and simplify a bit to see that the above
expression is
lim
h→0
15t
2
+ 30th + 15h
2
+ 7 − 15t
2
− 7
h
= lim
h→0
30th + 15h
2
h
= lim
h→0
(30t + 15h).
It’s particularly nice that the h gets canceled from the denominator in the
last step, since that’s what was giving us all the trouble. Now we can just put
h = 0 to see that
instantaneous velocity at time t = lim
h→0
(30t + 15h) = 30t.
So at time 0, the car’s velocity is 30 ×0 = 0 mph—the car is at rest. Half an
hour later, at time t = 1/2, its velocity is 30 ×1/2 = 15 mph. One hour after
the start time, the velocity is 30. In fact, the fact that the velocity is 30t at
time t tells us that the car gets faster and faster at the constant rate of 30
mph every hour. That is, the car is constantly accelerating at 30 miles per
hour per hour, or 30 miles per hour squared.
5.2.4 The graphical interpretation of velocity
It’s time to look at a graph of the situation. Suppose that f (t) again represents
the position of the car at time t. If we want the instantaneous velocity at a
particular time t, we need to pick a time u close to t. Let’s draw the graph
of y = f(t) and mark in the points (t, f(t)) and (u, f(u)) as well as the line
through them:
88 • Continuity and Differentiability
PSfrag replacements
(
a, b)
[
a, b]
(
a, b]
[
a, b)
(
a, ∞)
[
a, ∞)
(
−∞, b)
(
−∞, b]
(
−∞, ∞)
{
x : a < x < b}
{
x : a ≤ x ≤ b}
{
x : a < x ≤ b}
{
x : a ≤ x < b}
{
x : x ≥ a}
{
x : x > a}
{
x : x ≤ b}
{
x : x < b}
R
a
b
shadow
0
1
4
−
2
3
−
3
g(
x) = x
2
f(
x) = x
3
g(
x) = x
2
f(
x) = x
3
mirror (
y = x)
f
−
1
(x) =
3
√
x
y = h
(x)
y = h
−
1
(x)
y = (
x − 1)
2
−
1
x
Same height
−
x
Same length,
opposite signs
y = −
2x
−
2
1
y =
1
2
x − 1
2
−
1
y = 2
x
y = 10
x
y = 2
−
x
y = log
2
(
x)
4
3 units
mirror (
x-axis)
y = |
x|
y = |
log
2
(x)|
θ radians
θ units
30
◦
=
π
6
45
◦
=
π
4
60
◦
=
π
3
120
◦
=
2
π
3
135
◦
=
3
π
4
150
◦
=
5
π
6
90
◦
=
π
2
180
◦
= π
210
◦
=
7
π
6
225
◦
=
5
π
4
240
◦
=
4
π
3
270
◦
=
3
π
2
300
◦
=
5
π
3
315
◦
=
7
π
4
330
◦
=
11
π
6
0
◦
= 0 radians
θ
hypotenuse
opposite
adjacent
0 (
≡ 2π)
π
2
π
3
π
2
I
II
III
IV
θ
(
x, y)
x
y
r
7
π
6
reference angle
reference angle =
π
6
sin +
sin −
cos +
cos −
tan +
tan −
A
S
T
C
7
π
4
9
π
13
5
π
6
(this angle is
5
π
6
clockwise)
1
2
1
2
3
4
5
6
0
−
1
−
2
−
3
−
4
−
5
−
6
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
π
2
y = sin(
x)
1
0
−
1
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
5
π
2
2
π
2
π
3
π
2
π
π
2
y = sin(
x)
y = cos(
x)
−
π
2
π
2
y = tan(
x), −
π
2
< x <
π
2
0
−
π
2
π
2
y = tan(
x)
−
2π
−
3π
−
5
π
2
−
3
π
2
−
π
−
π
2
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
y = sec(
x)
y = csc(
x)
y = cot(
x)
y = f(
x)
−
1
1
2
y = g(
x)
3
y = h
(x)
4
5
−
2
f(
x) =
1
x
g(
x) =
1
x
2
etc.
0
1
π
1
2
π
1
3
π
1
4
π
1
5
π
1
6
π
1
7
π
g(
x) = sin
1
x
1
0
−
1
L
10
100
200
y =
π
2
y = −
π
2
y = tan
−
1
(x)
π
2
π
y =
sin(
x)
x
, x > 3
0
1
−
1
a
L
f(
x) = x sin (1/x)
(0 < x < 0
.3)
h
(x) = x
g(
x) = −x
a
L
lim
x→a
+
f(x) = L
lim
x→a
+
f(x) = ∞
lim
x→a
+
f(x) = −∞
lim
x→a
+
f(x) DNE
lim
x→a
−
f(x) = L
lim
x→a
−
f(x) = ∞
lim
x→a
−
f(x) = −∞
lim
x→a
−
f(x) DNE
M
}
lim
x→a
−
f(x) = M
lim
x→a
f(x) = L
lim
x→a
f(x) DNE
lim
x→∞
f(x) = L
lim
x→∞
f(x) = ∞
lim
x→∞
f(x) = −∞
lim
x→∞
f(x) DNE
lim
x→−∞
f(x) = L
lim
x→−∞
f(x) = ∞
lim
x→−∞
f(x) = −∞
lim
x→−∞
f(x) DNE
lim
x →a
+
f(x) = ∞
lim
x →a
+
f(x) = −∞
lim
x →a
−
f(x) = ∞
lim
x →a
−
f(x) = −∞
lim
x →a
f(x) = ∞
lim
x →a
f(x) = −∞
lim
x →a
f(x) DNE
y = f (x)
a
y =
|x|
x
1
−1
y =
|x + 2|
x + 2
1
−1
−2
1
2
3
4
a
a
b
y = x sin
1
x
y = x
y = −x
a
b
c
d
C
a
b
c
d
−1
0
1
2
3
time
y
t
u
(t, f(t))
(u, f(u))
The slope of this line is given by
slope =
f(u) − f(t)
u − t
,
which is exactly the formula for the average velocity v
t↔u
from the previous
section. So we have a graphical interpretation for average velocity over the
time period t to u: it’s the slope of the line joining the points (t, f(t)) and
(u, f(u)) on the graph of position versus time.
Let’s try to find a similar interpretation for the instantaneous velocity. We
need to take the limit as u goes to t, so let’s repeat the previous graph a few
times, each time with u closer and closer to the fixed value t:
PSfrag replacements
(
a, b)
[
a, b]
(
a, b]
[
a, b)
(
a, ∞)
[
a, ∞)
(
−∞, b)
(
−∞, b]
(
−∞, ∞)
{
x : a < x < b}
{
x : a ≤ x ≤ b}
{
x : a < x ≤ b}
{
x : a ≤ x < b}
{
x : x ≥ a}
{
x : x > a}
{
x : x ≤ b}
{
x : x < b}
R
a
b
shadow
0
1
4
−
2
3
−
3
g(
x) = x
2
f(
x) = x
3
g(
x) = x
2
f(
x) = x
3
mirror (
y = x)
f
−
1
(x) =
3
√
x
y = h
(x)
y = h
−
1
(x)
y = (
x − 1)
2
−
1
x
Same height
−
x
Same length,
opposite signs
y = −
2x
−
2
1
y =
1
2
x − 1
2
−
1
y = 2
x
y = 10
x
y = 2
−
x
y = log
2
(
x)
4
3 units
mirror (
x-axis)
y = |
x|
y = |
log
2
(x)|
θ radians
θ units
30
◦
=
π
6
45
◦
=
π
4
60
◦
=
π
3
120
◦
=
2
π
3
135
◦
=
3
π
4
150
◦
=
5
π
6
90
◦
=
π
2
180
◦
= π
210
◦
=
7
π
6
225
◦
=
5
π
4
240
◦
=
4
π
3
270
◦
=
3
π
2
300
◦
=
5
π
3
315
◦
=
7
π
4
330
◦
=
11
π
6
0
◦
= 0 radians
θ
hypotenuse
opposite
adjacent
0 (
≡ 2π)
π
2
π
3
π
2
I
II
III
IV
θ
(
x, y)
x
y
r
7
π
6
reference angle
reference angle =
π
6
sin +
sin −
cos +
cos −
tan +
tan −
A
S
T
C
7
π
4
9
π
13
5
π
6
(this angle is
5
π
6
clockwise)
1
2
1
2
3
4
5
6
0
−
1
−
2
−
3
−
4
−
5
−
6
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
π
2
y = sin(
x)
1
0
−
1
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
5
π
2
2
π
2
π
3
π
2
π
π
2
y = sin(
x)
y = cos(
x)
−
π
2
π
2
y = tan(
x), −
π
2
< x <
π
2
0
−
π
2
π
2
y = tan(
x)
−
2π
−
3π
−
5
π
2
−
3
π
2
−
π
−
π
2
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
y = sec(
x)
y = csc(
x)
y = cot(
x)
y = f(
x)
−
1
1
2
y = g(
x)
3
y = h(x)
4
5
−2
f(x) =
1
x
g(x) =
1
x
2
etc.
0
1
π
1
2π
1
3π
1
4π
1
5π
1
6π
1
7π
g(x) = sin
1
x
1
0
−1
L
10
100
200
y =
π
2
y = −
π
2
y = tan
−1
(x)
π
2π
y =
sin(x)
x
, x > 3
0
1
−1
a
L
f(x) = x sin (1/x)
(0 < x < 0.3)
h(x) = x
g(x) = −x
a
L
lim
x→a
+
f(x) = L
lim
x→a
+
f(x) = ∞
lim
x→a
+
f(x) = −∞
lim
x→a
+
f(x) DNE
lim
x→a
−
f(x) = L
lim
x→a
−
f(x) = ∞
lim
x→a
−
f(x) = −∞
lim
x→a
−
f(x) DNE
M
}
lim
x→a
−
f(x) = M
lim
x→a
f(x) = L
lim
x→a
f(x) DNE
lim
x→∞
f(x) = L
lim
x→∞
f(x) = ∞
lim
x→∞
f(x) = −∞
lim
x→∞
f(x) DNE
lim
x→−∞
f(x) = L
lim
x→−∞
f(x) = ∞
lim
x→−∞
f(x) = −∞
lim
x→−∞
f(x) DNE
lim
x →a
+
f(x) = ∞
lim
x →a
+
f(x) = −∞
lim
x →a
−
f(x) = ∞
lim
x →a
−
f(x) = −∞
lim
x →a
f(x) = ∞
lim
x →a
f(x) = −∞
lim
x →a
f(x) DNE
y = f(x)
a
y =
|x|
x
1
−1
y =
|x + 2|
x + 2
1
−1
−2
1
2
3
4
a
a
b
y = x sin
1
x
y = x
y = −x
a
b
c
d
C
a
b
c
d
−1
0
1
2
3
time
y
t
u
(t, f(t))
(u, f(u))
time
time
time
yyy
t
tt
u
u
u
The lines seem to be getting closer to the tangent line at the point (t, f (t)).
Since the instantaneous velocity is the limit of the slopes of these lines as
u → t, we’d like to say that the instantaneous velocity is exactly equal to the
slope of the tangent line through (t, f(t)). Looks like we need to understand
tangent lines better. . . .
5.2.5 Tangent lines
Suppose we pick a number x in the domain of some function f. Then the
point (x, f(x)) lies on the graph of y = f(x). We want to try to draw a line
through that point which is tangential to the curve—that is, we want to find a
tangent line. Intuitively, this means that the line we’re looking for just grazes
the curve at our point (x, f (x)). The tangent line doesn’t have to intersect
the curve only once! For example, the tangent line through (x, f(x)) in the
following picture hits the curve again, and that’s not a problem:
Section 5.2.5: Tangent lines • 89
PSfrag
replacements
(
a, b)
[
a, b]
(
a, b]
[
a, b)
(
a, ∞)
[
a, ∞)
(
−∞, b)
(
−∞, b]
(
−∞, ∞)
{
x : a < x < b}
{
x : a ≤ x ≤ b}
{
x : a < x ≤ b}
{
x : a ≤ x < b}
{
x : x ≥ a}
{
x : x > a}
{
x : x ≤ b}
{
x : x < b}
R
a
b
shado
w
0
1
4
−
2
3
−
3
g(
x) = x
2
f(
x) = x
3
g(
x) = x
2
f(
x) = x
3
mirror
(y = x)
f
−
1
(x) =
3
√
x
y = h
(x)
y = h
−
1
(x)
y =
(x − 1)
2
−
1
x
Same
height
−
x
Same
length,
opp
osite signs
y = −
2x
−
2
1
y =
1
2
x − 1
2
−
1
y =
2
x
y =
10
x
y =
2
−x
y =
log
2
(x)
4
3
units
mirror
(x-axis)
y = |
x|
y = |
log
2
(x)|
θ radians
θ units
30
◦
=
π
6
45
◦
=
π
4
60
◦
=
π
3
120
◦
=
2
π
3
135
◦
=
3
π
4
150
◦
=
5
π
6
90
◦
=
π
2
180
◦
= π
210
◦
=
7
π
6
225
◦
=
5
π
4
240
◦
=
4
π
3
270
◦
=
3
π
2
300
◦
=
5
π
3
315
◦
=
7
π
4
330
◦
=
11
π
6
0
◦
=
0 radians
θ
hypotenuse
opp
osite
adjacen
t
0
(≡ 2π)
π
2
π
3
π
2
I
I
I
I
II
IV
θ
(
x, y)
x
y
r
7
π
6
reference
angle
reference
angle =
π
6
sin
+
sin −
cos
+
cos −
tan
+
tan −
A
S
T
C
7
π
4
9
π
13
5
π
6
(this
angle is
5π
6
clo
ckwise)
1
2
1
2
3
4
5
6
0
−
1
−
2
−
3
−
4
−
5
−
6
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
π
2
y =
sin(x)
1
0
−
1
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
5
π
2
2
π
2
π
3
π
2
π
π
2
y =
sin(x)
y =
cos(x)
−
π
2
π
2
y =
tan(x), −
π
2
<
x <
π
2
0
−
π
2
π
2
y =
tan(x)
−
2π
−
3π
−
5
π
2
−
3
π
2
−
π
−
π
2
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
y =
sec(x)
y =
csc(x)
y =
cot(x)
y = f(
x)
−
1
1
2
y = g(
x)
3
y = h
(x)
4
5
−
2
f(
x) =
1
x
g(
x) =
1
x
2
etc.
0
1
π
1
2
π
1
3
π
1
4
π
1
5
π
1
6
π
1
7
π
g(
x) = sin
1
x
1
0
−
1
L
10
100
200
y =
π
2
y = −
π
2
y =
tan
−1
(x)
π
2
π
y =
sin
(x)
x
,
x > 3
0
1
−
1
a
L
f(
x) = x sin (1/x)
(0 <
x < 0.3)
h
(x) = x
g(
x) = −x
a
L
lim
x
→a
+
f(x) = L
lim
x
→a
+
f(x) = ∞
lim
x
→a
+
f(x) = −∞
lim
x
→a
+
f(x) DNE
lim
x
→a
−
f(x) = L
lim
x→a
−
f(x) = ∞
lim
x→a
−
f(x) = −∞
lim
x→a
−
f(x) DNE
M
}
lim
x→a
−
f(x) = M
lim
x→a
f(x) = L
lim
x→a
f(x) DNE
lim
x→∞
f(x) = L
lim
x→∞
f(x) = ∞
lim
x→∞
f(x) = −∞
lim
x→∞
f(x) DNE
lim
x→−∞
f(x) = L
lim
x→−∞
f(x) = ∞
lim
x→−∞
f(x) = −∞
lim
x→−∞
f(x) DNE
lim
x →a
+
f(x) = ∞
lim
x →a
+
f(x) = −∞
lim
x →a
−
f(x) = ∞
lim
x →a
−
f(x) = −∞
lim
x →a
f(x) = ∞
lim
x →a
f(x) = −∞
lim
x →a
f(x) DNE
y = f (x)
a
y =
|x|
x
1
−1
y =
|x + 2|
x + 2
1
−1
−2
1
2
3
4
a
a
b
y = x sin
1
x
y = x
y = −x
a
b
c
d
C
a
b
c
d
−1
0
1
2
3
time
y
t
u
(t, f(t))
(u, f(u))
time
y
t
u
y
x
(x, f(x))
It’s possible that there’s no tangent line through a given point on a graph.
For example, consider the graph of y = |x|:
PSfrag
replacements
(
a, b)
[
a, b]
(
a, b]
[
a, b)
(
a, ∞)
[
a, ∞)
(
−∞, b)
(
−∞, b]
(
−∞, ∞)
{
x : a < x < b}
{
x : a ≤ x ≤ b}
{
x : a < x ≤ b}
{
x : a ≤ x < b}
{
x : x ≥ a}
{
x : x > a}
{
x : x ≤ b}
{
x : x < b}
R
a
b
shado
w
0
1
4
−
2
3
−
3
g(
x) = x
2
f(
x) = x
3
g(
x) = x
2
f(
x) = x
3
mirror
(y = x)
f
−
1
(x) =
3
√
x
y = h
(x)
y = h
−
1
(x)
y =
(x − 1)
2
−
1
x
Same
height
−
x
Same
length,
opp
osite signs
y = −
2x
−
2
1
y =
1
2
x − 1
2
−
1
y =
2
x
y =
10
x
y =
2
−x
y =
log
2
(x)
4
3
units
mirror
(x-axis)
y = |
x|
y = |
log
2
(x)|
θ radians
θ units
30
◦
=
π
6
45
◦
=
π
4
60
◦
=
π
3
120
◦
=
2
π
3
135
◦
=
3
π
4
150
◦
=
5
π
6
90
◦
=
π
2
180
◦
= π
210
◦
=
7
π
6
225
◦
=
5
π
4
240
◦
=
4
π
3
270
◦
=
3
π
2
300
◦
=
5
π
3
315
◦
=
7
π
4
330
◦
=
11
π
6
0
◦
=
0 radians
θ
hypotenuse
opp
osite
adjacen
t
0
(≡ 2π)
π
2
π
3
π
2
I
I
I
I
II
IV
θ
(
x, y)
x
y
r
7
π
6
reference
angle
reference
angle =
π
6
sin
+
sin −
cos
+
cos −
tan
+
tan −
A
S
T
C
7
π
4
9
π
13
5
π
6
(this
angle is
5π
6
clo
ckwise)
1
2
1
2
3
4
5
6
0
−
1
−
2
−
3
−
4
−
5
−
6
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
π
2
y =
sin(x)
1
0
−
1
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
5
π
2
2
π
2
π
3
π
2
π
π
2
y =
sin(x)
y =
cos(x)
−
π
2
π
2
y =
tan(x), −
π
2
<
x <
π
2
0
−
π
2
π
2
y =
tan(x)
−
2π
−
3π
−
5
π
2
−
3
π
2
−
π
−
π
2
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
y =
sec(x)
y =
csc(x)
y =
cot(x)
y = f(
x)
−
1
1
2
y = g(
x)
3
y = h
(x)
4
5
−
2
f(
x) =
1
x
g(
x) =
1
x
2
etc.
0
1
π
1
2
π
1
3
π
1
4
π
1
5
π
1
6
π
1
7
π
g(
x) = sin
1
x
1
0
−
1
L
10
100
200
y =
π
2
y = −
π
2
y =
tan
−1
(x)
π
2
π
y =
sin(
x)
x
,
x > 3
0
1
−
1
a
L
f(
x) = x sin (1/x)
(0 < x < 0.3)
h(x) = x
g(x) = −x
a
L
lim
x→a
+
f(x) = L
lim
x→a
+
f(x) = ∞
lim
x→a
+
f(x) = −∞
lim
x→a
+
f(x) DNE
lim
x→a
−
f(x) = L
lim
x→a
−
f(x) = ∞
lim
x→a
−
f(x) = −∞
lim
x→a
−
f(x) DNE
M
}
lim
x→a
−
f(x) = M
lim
x→a
f(x) = L
lim
x→a
f(x) DNE
lim
x→∞
f(x) = L
lim
x→∞
f(x) = ∞
lim
x→∞
f(x) = −∞
lim
x→∞
f(x) DNE
lim
x→−∞
f(x) = L
lim
x→−∞
f(x) = ∞
lim
x→−∞
f(x) = −∞
lim
x→−∞
f(x) DNE
lim
x →a
+
f(x) = ∞
lim
x →a
+
f(x) = −∞
lim
x →a
−
f(x) = ∞
lim
x →a
−
f(x) = −∞
lim
x →a
f(x) = ∞
lim
x →a
f(x) = −∞
lim
x →a
f(x) DNE
y = f(x)
a
y =
|x|
x
1
−1
y =
|x + 2|
x + 2
1
−1
−2
1
2
3
4
a
a
b
y = x sin
1
x
y = x
y = −x
a
b
c
d
C
a
b
c
d
−1
0
1
2
3
time
y
t
u
(t, f(t))
(u, f(u))
time
y
t
u
y
x
(x, f(x))
y = |x|
The graph passes through (0, 0), but there’s no tangent line through that
point. What could the tangent line possibly be, after all? No matter what
you draw, you can’t cuddle up to the graph there since it’s got a sharp point
at the origin. We’ll return to this example in Section 5.2.10 below.
Even if the tangent line through (x, f(x)) exists, how on earth do you
find it? Remember, to specify a line, you only need to provide two pieces
of information: a point the line goes through and its slope. Then you can
use the point-slope form to find the equation of the line. Well, we have one
ingredient: we know the line passes through the point (x, f (x)). Now we just
need to find the slope. To do this, we’ll play a game similar to the one we
played with instantaneous velocities in the previous section.
Start by picking a number z which is close to x (either to the right or to
the left) and plot the point (z, f (z)) on the curve. Now draw the line through
the points (x, f(x)) and (z, f(z)):
PSfrag
replacements
(
a, b)
[
a, b]
(
a, b]
[
a, b)
(
a, ∞)
[
a, ∞)
(
−∞, b)
(
−∞, b]
(
−∞, ∞)
{
x : a < x < b}
{
x : a ≤ x ≤ b}
{
x : a < x ≤ b}
{
x : a ≤ x < b}
{
x : x ≥ a}
{
x : x > a}
{
x : x ≤ b}
{
x : x < b}
R
a
b
shado
w
0
1
4
−
2
3
−
3
g(
x) = x
2
f(
x) = x
3
g(
x) = x
2
f(
x) = x
3
mirror
(y = x)
f
−
1
(x) =
3
√
x
y = h
(x)
y = h
−
1
(x)
y =
(x − 1)
2
−
1
x
Same
height
−
x
Same
length,
opp
osite signs
y = −
2x
−
2
1
y =
1
2
x − 1
2
−
1
y =
2
x
y =
10
x
y =
2
−x
y =
log
2
(x)
4
3
units
mirror
(x-axis)
y = |
x|
y = |
log
2
(x)|
θ radians
θ units
30
◦
=
π
6
45
◦
=
π
4
60
◦
=
π
3
120
◦
=
2
π
3
135
◦
=
3
π
4
150
◦
=
5
π
6
90
◦
=
π
2
180
◦
= π
210
◦
=
7
π
6
225
◦
=
5
π
4
240
◦
=
4
π
3
270
◦
=
3
π
2
300
◦
=
5
π
3
315
◦
=
7
π
4
330
◦
=
11
π
6
0
◦
=
0 radians
θ
hypotenuse
opp
osite
adjacen
t
0
(≡ 2π)
π
2
π
3
π
2
I
I
I
I
II
IV
θ
(
x, y)
x
y
r
7
π
6
reference
angle
reference
angle =
π
6
sin
+
sin −
cos
+
cos −
tan
+
tan −
A
S
T
C
7
π
4
9
π
13
5
π
6
(this
angle is
5π
6
clo
ckwise)
1
2
1
2
3
4
5
6
0
−
1
−
2
−
3
−
4
−
5
−
6
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
π
2
y =
sin(x)
1
0
−
1
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
5
π
2
2
π
2
π
3
π
2
π
π
2
y =
sin(x)
y =
cos(x)
−
π
2
π
2
y =
tan(x), −
π
2
<
x <
π
2
0
−
π
2
π
2
y =
tan(x)
−
2π
−
3π
−
5
π
2
−
3
π
2
−
π
−
π
2
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
y =
sec(x)
y =
csc(x)
y =
cot(x)
y = f(
x)
−
1
1
2
y = g(
x)
3
y = h
(x)
4
5
−
2
f(
x) =
1
x
g(
x) =
1
x
2
etc.
0
1
π
1
2
π
1
3
π
1
4
π
1
5
π
1
6π
1
7π
g(x) = sin
1
x
1
0
−1
L
10
100
200
y =
π
2
y = −
π
2
y = tan
−1
(x)
π
2π
y =
sin(x)
x
, x > 3
0
1
−1
a
L
f(x) = x sin (1/x)
(0 < x < 0.3)
h(x) = x
g(x) = −x
a
L
lim
x→a
+
f(x) = L
lim
x→a
+
f(x) = ∞
lim
x→a
+
f(x) = −∞
lim
x→a
+
f(x) DNE
lim
x→a
−
f(x) = L
lim
x→a
−
f(x) = ∞
lim
x→a
−
f(x) = −∞
lim
x→a
−
f(x) DNE
M
}
lim
x→a
−
f(x) = M
lim
x→a
f(x) = L
lim
x→a
f(x) DNE
lim
x→∞
f(x) = L
lim
x→∞
f(x) = ∞
lim
x→∞
f(x) = −∞
lim
x→∞
f(x) DNE
lim
x→−∞
f(x) = L
lim
x→−∞
f(x) = ∞
lim
x→−∞
f(x) = −∞
lim
x→−∞
f(x) DNE
lim
x →a
+
f(x) = ∞
lim
x →a
+
f(x) = −∞
lim
x →a
−
f(x) = ∞
lim
x →a
−
f(x) = −∞
lim
x →a
f(x) = ∞
lim
x →a
f(x) = −∞
lim
x →a
f(x) DNE
y = f (x)
a
y =
|x|
x
1
−1
y =
|x + 2|
x + 2
1
−1
−2
1
2
3
4
a
a
b
y = x sin
1
x
y = x
y = −x
a
b
c
d
C
a
b
c
d
−1
0
1
2
3
time
y
t
u
(t, f(t))
(u, f(u))
time
y
t
u
y
x
(x, f(x))
y = |x|
(z, f(z))
z
y = f(x)
90 • Continuity and Differentiability
Since the slope is the rise over the run, the slope of the dashed line is
f(z) − f(x)
z − x
.
Now, as the point z gets closer and closer to x, without ever actually getting
to x itself, the slope of the above line should become closer and closer to the
slope of the tangent we’re looking for. So, if there’s any justice in the world,
then it should be true that
slope of tangent line through (x, f (x)) = lim
z→x
f(z) − f(x)
z − x
.
Let’s set h = z − x; then we see that as z → x, we have h → 0, so we also
have
slope of tangent line through (x, f(x)) = lim
h→0
f(x + h) − f(x)
h
.
Of course, this only makes sense if the limit actually exists!
5.2.6 The derivative function
In the following picture, I’ve drawn in the tangent lines through three different
points on the curve:
PSfrag
replacements
(
a, b)
[
a, b]
(
a, b]
[
a, b)
(
a, ∞)
[
a, ∞)
(
−∞, b)
(
−∞, b]
(
−∞, ∞)
{
x : a < x < b}
{
x : a ≤ x ≤ b}
{
x : a < x ≤ b}
{
x : a ≤ x < b}
{
x : x ≥ a}
{
x : x > a}
{
x : x ≤ b}
{
x : x < b}
R
a
b
shado
w
0
1
4
−
2
3
−
3
g(
x) = x
2
f(
x) = x
3
g(
x) = x
2
f(
x) = x
3
mirror
(y = x)
f
−
1
(x) =
3
√
x
y = h
(x)
y = h
−
1
(x)
y =
(x − 1)
2
−
1
x
Same
height
−
x
Same
length,
opp
osite signs
y = −
2x
−
2
1
y =
1
2
x − 1
2
−
1
y =
2
x
y =
10
x
y =
2
−x
y =
log
2
(x)
4
3
units
mirror
(x-axis)
y = |
x|
y = |
log
2
(x)|
θ radians
θ units
30
◦
=
π
6
45
◦
=
π
4
60
◦
=
π
3
120
◦
=
2
π
3
135
◦
=
3
π
4
150
◦
=
5
π
6
90
◦
=
π
2
180
◦
= π
210
◦
=
7
π
6
225
◦
=
5
π
4
240
◦
=
4
π
3
270
◦
=
3
π
2
300
◦
=
5
π
3
315
◦
=
7
π
4
330
◦
=
11
π
6
0
◦
=
0 radians
θ
hypotenuse
opp
osite
adjacen
t
0
(≡ 2π)
π
2
π
3
π
2
I
I
I
I
II
IV
θ
(
x, y)
x
y
r
7
π
6
reference
angle
reference
angle =
π
6
sin
+
sin −
cos
+
cos −
tan
+
tan −
A
S
T
C
7
π
4
9
π
13
5
π
6
(this
angle is
5π
6
clo
ckwise)
1
2
1
2
3
4
5
6
0
−
1
−
2
−
3
−
4
−
5
−
6
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
π
2
y =
sin(x)
1
0
−
1
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
5
π
2
2
π
2
π
3
π
2
π
π
2
y =
sin(x)
y =
cos(x)
−
π
2
π
2
y =
tan(x), −
π
2
<
x <
π
2
0
−
π
2
π
2
y =
tan(x)
−
2π
−
3π
−
5
π
2
−
3
π
2
−
π
−
π
2
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
y =
sec(x)
y =
csc(x)
y =
cot(x)
y = f(
x)
−
1
1
2
y = g(
x)
3
y = h
(x)
4
5
−
2
f(
x) =
1
x
g(
x) =
1
x
2
etc.
0
1
π
1
2
π
1
3
π
1
4
π
1
5
π
1
6
π
1
7
π
g(
x) = sin
1
x
1
0
−
1
L
10
100
200
y =
π
2
y = −
π
2
y =
tan
−1
(x)
π
2
π
y =
sin(
x)
x
,
x > 3
0
1
−
1
a
L
f(x) = x sin (1/x)
(0 < x < 0.3)
h(x) = x
g(x) = −x
a
L
lim
x→a
+
f(x) = L
lim
x→a
+
f(x) = ∞
lim
x→a
+
f(x) = −∞
lim
x→a
+
f(x) DNE
lim
x→a
−
f(x) = L
lim
x→a
−
f(x) = ∞
lim
x→a
−
f(x) = −∞
lim
x→a
−
f(x) DNE
M
}
lim
x→a
−
f(x) = M
lim
x→a
f(x) = L
lim
x→a
f(x) DNE
lim
x→∞
f(x) = L
lim
x→∞
f(x) = ∞
lim
x→∞
f(x) = −∞
lim
x→∞
f(x) DNE
lim
x→−∞
f(x) = L
lim
x→−∞
f(x) = ∞
lim
x→−∞
f(x) = −∞
lim
x→−∞
f(x) DNE
lim
x →a
+
f(x) = ∞
lim
x →a
+
f(x) = −∞
lim
x →a
−
f(x) = ∞
lim
x →a
−
f(x) = −∞
lim
x →a
f(x) = ∞
lim
x →a
f(x) = −∞
lim
x →a
f(x) DNE
y = f (x)
a
y =
|x|
x
1
−1
y =
|x + 2|
x + 2
1
−1
−2
1
2
3
4
a
a
b
y = x sin
1
x
y = x
y = −x
a
b
c
d
C
a
b
c
d
−1
0
1
2
3
time
y
t
u
(t, f(t))
(u, f(u))
time
y
t
u
y
x
(x, f(x))
y = |x|
(z, f(z))
z
y = f(x)
a
tangent at x = a
b
tangent at x = b
c
tangent at x = c
These lines have different slopes. That is, the slope of the tangent line de-
pends on which value of x you start with. Another way of saying this is that
the slope of the tangent line through (x, f(x)) is itself a function of x. This
function is called the derivative of f and is written as f
0
. We say that we have
differentiated the function f with respect to its variable x to get the function
f
0
. By the formula at the end of the previous section, we see that
f
0
(x) = lim
h→0
f(x + h) − f (x)
h
provided that the limit exists. In this case, we say that f is differentiable at x.
If the limit doesn’t exist for some particular x, then that value of x is not in
the domain of the derivative function f
0
, so we say that f is not differentiable
at x. The limit could fail to exist for a variety of reasons. In particular, there
Section 5.2.7: The derivative as a limiting ratio • 91
could be a sharp corner as in the example of y = |x| above. On an even more
basic level, if x isn’t in the domain of f, then you can’t even plot the point
(x, f(x)), let alone draw a tangent line there!
Now let’s recall the definition of instantaneous velocity in Section 5.2.3
above:
instantaneous velocity at time t = lim
h→0
f(t + h) − f(t)
h
,
where f(t) is the position of the car at time t. This right-hand side of this
the same as the definition of f
0
(x) above, except with x replaced by t! That
is, if v(t) is the instantaneous velocity at time t, then v(t) = f
0
(t). Velocity
is precisely the derivative of position with respect to time.
Let’s look at one example of finding a derivative. If f(x) = x
2
, what
is f
0
(x)? The computation is very similar to the one we did at the end of
PSfrag
replacements
(
a, b)
[
a, b]
(
a, b]
[
a, b)
(
a, ∞)
[
a, ∞)
(
−∞, b)
(
−∞, b]
(
−∞, ∞)
{
x : a < x < b}
{
x : a ≤ x ≤ b}
{
x : a < x ≤ b}
{
x : a ≤ x < b}
{
x : x ≥ a}
{
x : x > a}
{
x : x ≤ b}
{
x : x < b}
R
a
b
shado
w
0
1
4
−
2
3
−
3
g(
x) = x
2
f(
x) = x
3
g(
x) = x
2
f(
x) = x
3
mirror
(y = x)
f
−
1
(x) =
3
√
x
y = h
(x)
y = h
−
1
(x)
y =
(x − 1)
2
−
1
x
Same
height
−
x
Same
length,
opp
osite signs
y = −
2x
−
2
1
y =
1
2
x − 1
2
−
1
y =
2
x
y =
10
x
y =
2
−x
y =
log
2
(x)
4
3
units
mirror
(x-axis)
y = |
x|
y = |
log
2
(x)|
θ radians
θ units
30
◦
=
π
6
45
◦
=
π
4
60
◦
=
π
3
120
◦
=
2
π
3
135
◦
=
3
π
4
150
◦
=
5
π
6
90
◦
=
π
2
180
◦
= π
210
◦
=
7
π
6
225
◦
=
5
π
4
240
◦
=
4
π
3
270
◦
=
3
π
2
300
◦
=
5
π
3
315
◦
=
7
π
4
330
◦
=
11
π
6
0
◦
=
0 radians
θ
hypotenuse
opp
osite
adjacen
t
0
(≡ 2π)
π
2
π
3
π
2
I
I
I
I
II
IV
θ
(
x, y)
x
y
r
7
π
6
reference
angle
reference
angle =
π
6
sin
+
sin −
cos
+
cos −
tan
+
tan −
A
S
T
C
7
π
4
9
π
13
5
π
6
(this
angle is
5π
6
clo
ckwise)
1
2
1
2
3
4
5
6
0
−
1
−
2
−
3
−
4
−
5
−
6
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
π
2
y =
sin(x)
1
0
−
1
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
5
π
2
2
π
2
π
3
π
2
π
π
2
y =
sin(x)
y =
cos(x)
−
π
2
π
2
y =
tan(x), −
π
2
<
x <
π
2
0
−
π
2
π
2
y =
tan(x)
−
2π
−
3π
−
5
π
2
−
3
π
2
−
π
−
π
2
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
y =
sec(x)
y =
csc(x)
y =
cot(x)
y = f(
x)
−
1
1
2
y = g(
x)
3
y = h
(x)
4
5
−
2
f(
x) =
1
x
g(
x) =
1
x
2
etc.
0
1
π
1
2
π
1
3
π
1
4
π
1
5
π
1
6
π
1
7
π
g(
x) = sin
1
x
1
0
−
1
L
10
100
200
y =
π
2
y = −
π
2
y =
tan
−1
(x)
π
2
π
y =
sin(
x)
x
,
x > 3
0
1
−
1
a
L
f(
x) = x sin (1/x)
(0 <
x < 0.3)
h
(x) = x
g(
x) = −x
a
L
lim
x
→a
+
f(x) = L
lim
x
→a
+
f(x) = ∞
lim
x
→a
+
f(x) = −∞
lim
x
→a
+
f(x) DNE
lim
x
→a
−
f(x) = L
lim
x
→a
−
f(x) = ∞
lim
x
→a
−
f(x) = −∞
lim
x
→a
−
f(x) DNE
M
}
lim
x→a
−
f(x) = M
lim
x→a
f(x) = L
lim
x→a
f(x) DNE
lim
x→∞
f(x) = L
lim
x→∞
f(x) = ∞
lim
x→∞
f(x) = −∞
lim
x→∞
f(x) DNE
lim
x→−∞
f(x) = L
lim
x→−∞
f(x) = ∞
lim
x→−∞
f(x) = −∞
lim
x→−∞
f(x) DNE
lim
x →a
+
f(x) = ∞
lim
x →a
+
f(x) = −∞
lim
x →a
−
f(x) = ∞
lim
x →a
−
f(x) = −∞
lim
x →a
f(x) = ∞
lim
x →a
f(x) = −∞
lim
x →a
f(x) DNE
y = f (x)
a
y =
|x|
x
1
−1
y =
|x + 2|
x + 2
1
−1
−2
1
2
3
4
a
a
b
y = x sin
1
x
y = x
y = −x
a
b
c
d
C
a
b
c
d
−1
0
1
2
3
time
y
t
u
(t, f(t))
(u, f(u))
time
y
t
u
y
x
(x, f(x))
y = |x|
(z, f(z))
z
y = f(x)
a
tangent at x = a
b
tangent at x = b
c
tangent at x = c
Section 5.2.3 above:
f
0
(x) = lim
h→0
f(x + h) − f(x)
h
= lim
h→0
(x + h)
2
− x
2
h
= lim
h→0
x
2
+ 2xh + h
2
− x
2
h
= lim
h→0
2xh + h
2
h
= lim
h→0
(2x + h) = 2x.
So the derivative of f(x) = x
2
is given by f
0
(x) = 2x. This means that the
slope of the tangent to the parabola y = x
2
at the point (x, x
2
) is precisely
2x. Let’s draw the curve and a few tangent lines to check it out:
&
PSfrag
replacements
(
a, b)
[
a, b]
(
a, b]
[
a, b)
(
a, ∞)
[
a, ∞)
(
−∞, b)
(
−∞, b]
(
−∞, ∞)
{
x : a < x < b}
{
x : a ≤ x ≤ b}
{
x : a < x ≤ b}
{
x : a ≤ x < b}
{
x : x ≥ a}
{
x : x > a}
{
x : x ≤ b}
{
x : x < b}
R
a
b
shado
w
0
1
4
−
2
3
−
3
g(
x) = x
2
f(
x) = x
3
g(
x) = x
2
f(
x) = x
3
mirror
(y = x)
f
−
1
(x) =
3
√
x
y = h
(x)
y = h
−
1
(x)
y =
(x − 1)
2
−
1
x
Same
height
−
x
Same
length,
opp
osite signs
y = −
2x
−
2
1
y =
1
2
x − 1
2
−
1
y =
2
x
y =
10
x
y =
2
−x
y =
log
2
(x)
4
3
units
mirror
(x-axis)
y = |
x|
y = |
log
2
(x)|
θ radians
θ units
30
◦
=
π
6
45
◦
=
π
4
60
◦
=
π
3
120
◦
=
2
π
3
135
◦
=
3
π
4
150
◦
=
5
π
6
90
◦
=
π
2
180
◦
= π
210
◦
=
7
π
6
225
◦
=
5
π
4
240
◦
=
4
π
3
270
◦
=
3
π
2
300
◦
=
5
π
3
315
◦
=
7
π
4
330
◦
=
11
π
6
0
◦
=
0 radians
θ
hypotenuse
opp
osite
adjacen
t
0
(≡ 2π)
π
2
π
3
π
2
I
I
I
I
II
IV
θ
(
x, y)
x
y
r
7
π
6
reference
angle
reference
angle =
π
6
sin
+
sin −
cos
+
cos −
tan
+
tan −
A
S
T
C
7
π
4
9
π
13
5
π
6
(this
angle is
5π
6
clo
ckwise)
1
2
1
2
3
4
5
6
0
−
1
−
2
−
3
−
4
−
5
−
6
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
π
2
y =
sin(x)
1
0
−
1
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
5
π
2
2
π
2
π
3
π
2
π
π
2
y =
sin(x)
y =
cos(x)
−
π
2
π
2
y =
tan(x), −
π
2
<
x <
π
2
0
−
π
2
π
2
y =
tan(x)
−
2π
−
3π
−
5
π
2
−
3
π
2
−
π
−
π
2
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
y =
sec(x)
y =
csc(x)
y =
cot(x)
y = f(
x)
−
1
1
2
y = g(
x)
3
y = h
(x)
4
5
−
2
f(
x) =
1
x
g(
x) =
1
x
2
etc.
0
1
π
1
2
π
1
3
π
1
4
π
1
5
π
1
6
π
1
7
π
g(
x) = sin
1
x
1
0
−
1
L
10
100
200
y =
π
2
y = −
π
2
y =
tan
−1
(x)
π
2
π
y =
sin(
x)
x
,
x > 3
0
1
−
1
a
L
f(x) = x sin (1/x)
(0 < x < 0.3)
h(x) = x
g(x) = −x
a
L
lim
x→a
+
f(x) = L
lim
x→a
+
f(x) = ∞
lim
x→a
+
f(x) = −∞
lim
x→a
+
f(x) DNE
lim
x→a
−
f(x) = L
lim
x→a
−
f(x) = ∞
lim
x→a
−
f(x) = −∞
lim
x→a
−
f(x) DNE
M
}
lim
x→a
−
f(x) = M
lim
x→a
f(x) = L
lim
x→a
f(x) DNE
lim
x→∞
f(x) = L
lim
x→∞
f(x) = ∞
lim
x→∞
f(x) = −∞
lim
x→∞
f(x) DNE
lim
x→−∞
f(x) = L
lim
x→−∞
f(x) = ∞
lim
x→−∞
f(x) = −∞
lim
x→−∞
f(x) DNE
lim
x →a
+
f(x) = ∞
lim
x →a
+
f(x) = −∞
lim
x →a
−
f(x) = ∞
lim
x →a
−
f(x) = −∞
lim
x →a
f(x) = ∞
lim
x →a
f(x) = −∞
lim
x →a
f(x) DNE
y = f(x)
a
y =
|x|
x
1
−1
y =
|x + 2|
x + 2
1
−1
−2
1
2
3
4
a
a
b
y = x sin
1
x
y = x
y = −x
a
b
c
d
C
a
b
c
d
−1
0
1
2
3
time
y
t
u
(t, f(t))
(u, f(u))
time
y
t
u
y
x
(x, f(x))
y = |x|
(z, f(z))
z
y = f(x)
a
tangent at x = a
b
tangent at x = b
c
tangent at x = c
y = x
2
tangent
at x = −1
The slope of the tangent at x = −1 does indeed look like it’s about −2, which
is consistent with the formula f
0
(x) = 2x. (Twice −1 is −2!) The same
is true with the other tangents—their slopes are all twice the corresponding
x-coordinate.
5.2.7 The derivative as a limiting ratio
In our formula for the derivative f
0
(x), we have to evaluate the quantity
f(x + h). What is this quantity? Well, if y = f(x), and you change x into
x + h, then f(x + h) is simply the new value of y. The amount h represents
how much you changed x, so let’s replace it by the quantity ∆x. Here the
symbol ∆ means “change in,” so that ∆x is just the change in x. (Don’t think
92 • Continuity and Differentiability
of ∆x as the product of ∆ and x—this is just plain wrong!) So, let’s rewrite
the formula for f
0
(x) with h replaced by ∆x:
f
0
(x) = lim
∆x→0
f(x + ∆x) − f (x)
∆x
.
OK, here’s what happens. We start out with our pair (x, y), where y = f (x).
We now take a new value of x, which we’ll call x
new
. The value of y then
changes as well to a new value y
new
, which of course is just f(x
new
). Now, the
amount of change of any quantity is just the new value minus the old one, so
we have two equations:
∆x = x
new
− x and ∆y = y
new
− y.
The first equation says that x
new
= x + ∆x, so now the second equation can
be transformed as follows:
∆y = y
new
− y = f (x
new
) − f(x) = f (x + ∆x) − f(x).
But this is just the numerator of the fraction in the definition of f
0
(x) above!
What this means is that
f
0
(x) = lim
∆x→0
∆y
∆x
.
An interpretation of this is that a small change in x produces approximately
f
0
(x) times as much change in y. Indeed, if y = f(x) = x
2
, then we’ve seen
in the previous section that f
0
(x) = 2x. Let’s concentrate on what happens
when x = 6, for example. First, note that our formula for f
0
(x) shows us that
f
0
(6) = 2 × 6 = 12. So, if you take the equation 6
2
= 36 and change the 6 a
little bit, the 36 will change by about 12 times as much. For example, if we
add 0.01 to 6, we should add 0.12 to 36. So I’m saying that (6.01)
2
should be
about 36.12. In fact, the actual answer is 36.1201, so I was really close.
Now, why didn’t I get the exact answer? The reason is that f
0
(x) isn’t
actually equal to the ratio of ∆y to ∆x: it’s equal to the limit of that ratio
as ∆x tends to 0. This means that if we don’t move as far away from 6, we
should do even better. Let’s try to guess the value of (6.0004)
2
. We have
changed our original x-value 6 by 0.0004, so the y-value should change by
12 times this much, which is 0.0048. Our guess is therefore that (6.0004)
2
is approximately 36.0048. Not bad—the actual answer is 36.00480016, so we
were very close! The smaller the change from 6, the better our method will
work.
Of course, the magic number 12 only works when you start at x = 6.
If instead you start at x = 13, the magic number is f
0
(13), which equals
2 × 13 = 26. So, we know 13
2
= 169; what is (13.0002)
2
? To get from 13
to 13.0002, you have to add 0.0002; since the magic number is 26, we have
to add 26 times as much to 169 to get our guess. That is, we add 0.0052 to
169 and come up with the guess 169.0052. Again, that’s pretty darn good:
(13.0002)
2
is actually exactly 169.00520004.
Anyway, we’ll return to these ideas in Chapter 13 when we look at lin-
earization. For now, let’s look at the formula
f
0
(x) = lim
∆x→0
∆y
∆x
.
Section 5.2.8: The derivative of linear functions • 93
once again. The right-hand side is the limit of the ratio of the change in y
to the change in x, as the change in x becomes small. Suppose that x is
so small that the change is barely noticeable. Instead of writing ∆x, which
means “change in x,” we’d now like to write dx, which should mean “really
really tiny change in x,” and similarly for y. Unfortunately neither dx nor dy
really means anything by itself;
∗
nevertheless this provides the inspiration for
writing the derivative in a different, more convenient way:
if y = f(x), then you can write
dy
dx
instead of f
0
(x).
For example, if y = x
2
, then
dy
dx
= 2x. In fact, if you replace y by x
2
, you get
a variety of different ways of expressing the same thing:
f
0
(x) =
dy
dx
=
d(x
2
)
dx
=
d
dx
(x
2
) = 2x.
As another example, in Section 5.2.3 above, we saw that if the position of a
car at time t is f (t) = 15t
2
+ 7, then its velocity is 30t. Remembering that
velocity is just f
0
(t), this means that f
0
(t) = 30t. If instead we decided to
call the position p, so that p = 15t
2
+ 7, we could write
dp
dt
= 30t. The point
is that not everything comes in x’s and y’s—you have to be able to deal with
other letters.
In summary, the quantity
dy
dx
is the derivative of y with respect to x. If
y = f (x), then
dy
dx
and f
0
(x) are the same thing. Finally, remember that the
quantity
dy
dx
is not actually a fraction at all—it’s the limit of the fraction
∆y
∆x
as ∆x → 0.
5.2.8 The derivative of linear functions
Let’s just pause for breath and go back to a simple case: suppose that f is
linear. This means that f(x) = mx + b for some m and b. What do you think
that f
0
(x) should be? Remember, this measures the slope of the tangent to
the curve y = f(x) at the point (x, f (x)). In our case, the graph of y = mx+b
is just a line of slope m and y-intercept equal to b. If there’s any justice in the
world, then the tangent at any point on the line is just the line itself! This
means that the value of f
0
(x) should be m no matter what x is: the curve
y = mx + b has constant slope m. Let’s check it out using the formula:
f
0
(x) = lim
h→0
f(x + h) − f (x)
h
= lim
h→0
(m(x + h) + b) − (mx + b)
h
= lim
h→0
mh
h
= lim
h→0
m = m.
So there is justice in the world: f
0
(x) = m regardless of what x is. That
is, the derivative of a linear function is constant. As you might expect, only
linear functions have constant slope (this is a consequence of the so-called
Mean Value Theorem; see Section 11.3.1 in Chapter 11). By the way, if f is
actually constant, so that f(x) = b, then the slope is always 0. In particular,
f
0
(x) = 0 for all x. So we’ve proved that the derivative of a constant function
is identically 0.
∗
There is a theory of “infinitesimals,” but it’s beyond the scope of this book!
94 • Continuity and Differentiability
5.2.9 Second and higher-order derivatives
Since you can start with a function f and take its derivative to get a new
function f
0
, you can actually take this new function and differentiate it again.
You end up with the derivative of the derivative; this is called the second
derivative, and it’s written as f
00
.
For example, we’ve seen that if f(x) = x
2
, then the derivative f
0
(x) = 2x.
PSfrag
replacements
(
a, b)
[
a, b]
(
a, b]
[
a, b)
(
a, ∞)
[
a, ∞)
(
−∞, b)
(
−∞, b]
(
−∞, ∞)
{
x : a < x < b}
{
x : a ≤ x ≤ b}
{
x : a < x ≤ b}
{
x : a ≤ x < b}
{
x : x ≥ a}
{
x : x > a}
{
x : x ≤ b}
{
x : x < b}
R
a
b
shado
w
0
1
4
−
2
3
−
3
g(
x) = x
2
f(
x) = x
3
g(
x) = x
2
f(
x) = x
3
mirror
(y = x)
f
−
1
(x) =
3
√
x
y = h
(x)
y = h
−
1
(x)
y =
(x − 1)
2
−
1
x
Same
height
−
x
Same
length,
opp
osite signs
y = −
2x
−
2
1
y =
1
2
x − 1
2
−
1
y =
2
x
y =
10
x
y =
2
−x
y =
log
2
(x)
4
3
units
mirror
(x-axis)
y = |
x|
y = |
log
2
(x)|
θ radians
θ units
30
◦
=
π
6
45
◦
=
π
4
60
◦
=
π
3
120
◦
=
2
π
3
135
◦
=
3
π
4
150
◦
=
5
π
6
90
◦
=
π
2
180
◦
= π
210
◦
=
7
π
6
225
◦
=
5
π
4
240
◦
=
4
π
3
270
◦
=
3
π
2
300
◦
=
5
π
3
315
◦
=
7
π
4
330
◦
=
11
π
6
0
◦
=
0 radians
θ
hypotenuse
opp
osite
adjacen
t
0
(≡ 2π)
π
2
π
3
π
2
I
I
I
I
II
IV
θ
(
x, y)
x
y
r
7
π
6
reference
angle
reference
angle =
π
6
sin
+
sin −
cos
+
cos −
tan
+
tan −
A
S
T
C
7
π
4
9
π
13
5
π
6
(this
angle is
5π
6
clo
ckwise)
1
2
1
2
3
4
5
6
0
−
1
−
2
−
3
−
4
−
5
−
6
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
π
2
y =
sin(x)
1
0
−
1
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
5
π
2
2
π
2
π
3
π
2
π
π
2
y =
sin(x)
y =
cos(x)
−
π
2
π
2
y =
tan(x), −
π
2
<
x <
π
2
0
−
π
2
π
2
y =
tan(x)
−
2π
−
3π
−
5
π
2
−
3
π
2
−
π
−
π
2
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
y =
sec(x)
y =
csc(x)
y =
cot(x)
y = f(
x)
−
1
1
2
y = g(
x)
3
y = h
(x)
4
5
−
2
f(
x) =
1
x
g(
x) =
1
x
2
etc.
0
1
π
1
2
π
1
3
π
1
4
π
1
5
π
1
6
π
1
7
π
g(
x) = sin
1
x
1
0
−
1
L
10
100
200
y =
π
2
y = −
π
2
y =
tan
−1
(x)
π
2
π
y =
sin(
x)
x
,
x > 3
0
1
−
1
a
L
f(
x) = x sin (1/x)
(0 <
x < 0.3)
h
(x) = x
g(
x) = −x
a
L
lim
x
→a
+
f(x) = L
lim
x
→a
+
f(x) = ∞
lim
x
→a
+
f(x) = −∞
lim
x
→a
+
f(x) DNE
lim
x
→a
−
f(x) = L
lim
x
→a
−
f(x) = ∞
lim
x
→a
−
f(x) = −∞
lim
x
→a
−
f(x) DNE
M
}
lim
x
→a
−
f(x) = M
lim
x
→a
f(x) = L
lim
x
→a
f(x) DNE
lim
x
→∞
f(x) = L
lim
x
→∞
f(x) = ∞
lim
x
→∞
f(x) = −∞
lim
x→∞
f(x) DNE
lim
x→−∞
f(x) = L
lim
x→−∞
f(x) = ∞
lim
x→−∞
f(x) = −∞
lim
x→−∞
f(x) DNE
lim
x →a
+
f(x) = ∞
lim
x →a
+
f(x) = −∞
lim
x →a
−
f(x) = ∞
lim
x →a
−
f(x) = −∞
lim
x →a
f(x) = ∞
lim
x →a
f(x) = −∞
lim
x →a
f(x) DNE
y = f (x)
a
y =
|x|
x
1
−1
y =
|x + 2|
x + 2
1
−1
−2
1
2
3
4
a
a
b
y = x sin
1
x
y = x
y = −x
a
b
c
d
C
a
b
c
d
−1
0
1
2
3
time
y
t
u
(t, f(t))
(u, f(u))
time
y
t
u
y
x
(x, f(x))
y = |x|
(z, f(z))
z
y = f(x)
a
tangent at x = a
b
tangent at x = b
c
tangent at x = c
y = x
2
tangent
at x = −1
Now we want to differentiate this result. Let’s put g(x) = 2x and try to work
out g
0
(x). Since g is a linear function with slope 2, we know from the previous
section above that g
0
(x) = 2. So the derivative of the derivative of f is the
constant function 2, and we have shown that f
00
(x) = 2 for all x.
If y = f(x), then we’ve seen that we can write
dy
dx
instead of f
0
(x). There’s
a similar sort of notation for the second derivative:
if y = f(x), then you can write
d
2
y
dx
2
instead of f
00
(x).
In the above example, if y = f(x) = x
2
, then we’ve seen that
f
00
(x) =
d
2
y
dx
2
=
d
2
(x
2
)
dx
2
=
d
2
dx
2
(x
2
) = 2.
These are all valid ways of expressing that the second derivative of f(x) = x
2
(with respect to x) is the constant function 2.
Why stop at taking two derivatives? The third derivative of a function
f is the derivative of the derivative of the derivative of f. That’s a lot of
derivatives! Realistically, you should think of the third derivative of f as
being the derivative of the second derivative of f, and you can write it in any
of the following ways:
f
000
(x), f
(3)
(x),
d
3
y
dx
3
, or
d
3
dx
3
(y).
The notation f
(3)
(x) is particularly convenient for higher derivatives, because
writing so many apostrophes is just plain stupid. So, for example, the fourth
derivative, which is just the derivative of the third derivative, would be written
f
(4)
(x) and not f
0000
(x). That said, it will sometimes be convenient to go the
other way and write f
(2)
(x) for the second derivative instead of f
00
(x). It’s
even possible to write f
(1)
(x) instead of f
0
(x), since we are only taking one
derivative, and also f
(0)
(x) instead of just f(x) itself (no derivatives!). That
way, any derivative can be written in the form f
(n)
(x) for some integer n.
5.2.10 When the derivative does not exist
In Section 5.2.5 above, I mentioned that the graph of f (x) = |x| has a sharp
corner at the origin. This should mean that the derivative doesn’t exist at
x = 0. Now let’s try to see why this is. Using the formula for the derivative,
we have
f
0
(x) = lim
h→0
f(x + h) − f(x)
h
= lim
h→0
|x + h| − |x|
h
.
We are interested in what happens when x = 0, so let’s replace x by 0 in the
above chain of equations:
f
0
(0) = lim
h→0
f(0 + h) − f(0)
h
= lim
h→0
|0 + h| − |0|
h
= lim
h→0
|h|
h
.
Section 5.2.10: When the derivative does not exist • 95
We have seen this limit before! In fact, in Section 4.6 of the previous chapter,
we saw that the limit does not exist. This means that the value of f
0
(0) is
undefined: 0 is not in the domain of f
0
. We also saw, however, that the above
limit does exist if you change it from a two-sided limit to a one-sided limit.
In particular, the right-hand limit is 1 and the left-hand limit is −1. This
motivates the idea of right-hand and left-hand derivatives, which are defined
by the formulas
lim
h→0
+
f(x + h) − f (x)
h
and lim
h→0
−
f(x + h) − f(x)
h
,
respectively. They look pretty similar to the definition of the ordinary deriva-
tive, except that the two-sided limit (that is, as h → 0) is replaced by right-
hand and left-hand limits, respectively. Just as in the case of limits, if the left-
and right-hand derivatives both exist and are equal, then the actual derivative
exists and is equal to the same thing. Also,
∗
if the derivative exists then the
left- and right-hand derivatives both exist and are equal to the derivative.
Anyway, the point is that if f(x) = |x|, at x = 0 the right-hand derivative
is 1 and the left-hand derivative is −1. Do you believe this? Look at the
graph again:
PSfrag replacements
(
a, b)
[
a, b]
(
a, b]
[
a, b)
(
a, ∞)
[
a, ∞)
(
−∞, b)
(
−∞, b]
(
−∞, ∞)
{
x : a < x < b}
{
x : a ≤ x ≤ b}
{
x : a < x ≤ b}
{
x : a ≤ x < b}
{
x : x ≥ a}
{
x : x > a}
{
x : x ≤ b}
{
x : x < b}
R
a
b
shadow
0
1
4
−
2
3
−
3
g(
x) = x
2
f(
x) = x
3
g(
x) = x
2
f(
x) = x
3
mirror (
y = x)
f
−
1
(x) =
3
√
x
y = h
(x)
y = h
−
1
(x)
y = (
x − 1)
2
−
1
x
Same height
−
x
Same length,
opposite signs
y = −
2x
−
2
1
y =
1
2
x − 1
2
−
1
y = 2
x
y = 10
x
y = 2
−
x
y = log
2
(
x)
4
3 units
mirror (
x-axis)
y = |
x|
y = |
log
2
(x)|
θ radians
θ units
30
◦
=
π
6
45
◦
=
π
4
60
◦
=
π
3
120
◦
=
2
π
3
135
◦
=
3
π
4
150
◦
=
5
π
6
90
◦
=
π
2
180
◦
= π
210
◦
=
7
π
6
225
◦
=
5
π
4
240
◦
=
4
π
3
270
◦
=
3
π
2
300
◦
=
5
π
3
315
◦
=
7
π
4
330
◦
=
11
π
6
0
◦
= 0 radians
θ
hypotenuse
opposite
adjacent
0 (
≡ 2π)
π
2
π
3
π
2
I
II
III
IV
θ
(
x, y)
x
y
r
7
π
6
reference angle
reference angle =
π
6
sin +
sin −
cos +
cos −
tan +
tan −
A
S
T
C
7
π
4
9
π
13
5
π
6
(this angle is
5
π
6
clockwise)
1
2
1
2
3
4
5
6
0
−
1
−
2
−
3
−
4
−
5
−
6
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
π
2
y = sin(
x)
1
0
−
1
−
3π
−
5
π
2
−
2π
−
3
π
2
−
π
−
π
2
3
π
5
π
2
2
π
2
π
3
π
2
π
π
2
y = sin(
x)
y = cos(
x)
−
π
2
π
2
y = tan(
x), −
π
2
< x <
π
2
0
−
π
2
π
2
y = tan(
x)
−
2π
−
3π
−
5
π
2
−
3
π
2
−
π
−
π
2
π
2
3
π
3
π
5
π
2
2
π
3
π
2
π
y = sec(
x)
y = csc(
x)
y = cot(
x)
y = f(
x)
−
1
1
2
y = g(
x)
3
y = h
(x)
4
5
−
2
f(
x) =
1
x
g(
x) =
1
x
2
etc.
0
1
π
1
2
π
1
3
π
1
4
π
1
5
π
1
6
π
1
7
π
g(
x) = sin
1
x
1
0
−
1
L
10
100
200
y =
π
2
y = −
π
2
y = tan
−
1
(x)
π
2
π
y =
sin(
x)
x
, x > 3
0
1
−
1
a
L
f(
x) = x sin (1/x)
(0 < x < 0
.3)
h
(x) = x
g(
x) = −x
a
L
lim
x→a
+
f(x) = L
lim
x→a
+
f(x) = ∞
lim
x→a
+
f(x) = −∞
lim
x→a
+
f(x) DNE
lim
x→a
−
f(x) = L
lim
x→a
−
f(x) = ∞
lim
x→a
−
f(x) = −∞
lim
x→a
−
f(x) DNE
M
}
lim
x→a
−
f(x) = M
lim
x→a
f(x) = L
lim
x→a
f(x) DNE
lim
x→∞
f(x) = L
lim
x→∞
f(x) = ∞
lim
x→∞
f(x) = −∞
lim
x→∞
f(x) DNE
lim
x→−∞
f(x) = L
lim
x→−∞
f(x) = ∞
lim
x→−∞
f(x) = −∞
lim
x→−∞
f(x) DNE
lim
x →a
+
f(x) = ∞
lim
x →a
+
f(x) = −∞
lim
x →a
−
f(x) = ∞
lim
x →a
−
f(x) = −∞
lim
x →a
f(x) = ∞
lim
x →a
f(x) = −∞
lim
x →a
f(x) DNE
y = f(x)
a
y =
|x|
x
1
−1
y =
|x + 2|
x + 2
1
−1
−2
1
2
3
4
a
a
b
y = x sin
1
x
y = x
y = −x
a
b
c
d
C
a
b
c
d
−1
0
1
2
3
time
y
t
u
(t, f(t))
(u, f(u))
time
y
t
u
y
x
(x, f(x))
y = |x|
(z, f(z))
z
y = f(x)
a
tangent at x = a
b
tangent at x = b
c
tangent at x = c
y = x
2
tangent
at x = −1
As you head from the origin along the curve to the right, it’s definitely slope
1 (and in fact it stays at slope 1, that is, f
0
(x) = 1 if x > 0). Similarly, to the
immediate left of the origin, the slope is −1 (and in fact f
0
(x) = −1 if x < 0).
Since the left-hand slope doesn’t equal the right-hand slope, there can be no
derivative at x = 0.
OK, so we have come up with a continuous function which isn’t differ-
entiable everywhere in its domain. Still, it is clearly differentiable except at
one measly little point. It turns out that you can have a continuous function
which is so spiky and jittery that it effectively has a sharp corner at every
single x, so it can’t be differentiated anywhere! This sort of funky function is
beyond the scope of this book, but I might as well mention that some of these
sorts of functions are used to model stock prices—if you’ve ever seen the graph
of a stock price, you’ll know what I mean by “spiky and jittery.” Anyway,
my point is that there are continuous functions which are not differentiable.
Are there any differentiable functions which aren’t continuous? The answer
is “no,” and we’re about to see why.
∗
You might say “conversely,” but only if you know what a “converse” is!