Section 9: Presentation of the Theory

_
_
toc
9. Presentation of the Theory
In this section, we present the demonstration of many of the theorems
discussed in the tutorial.
Theorem 9.1. Let a and c be numbers, then
lim
xa
c = c. (Rule 1)
Proof : Let > 0. In reference to the denition of limit, the function
under consideration is f(x) = c, and L = c. We want to choose a
number > 0 such that
0 < |x a| < = |f(x) L| <
or,
0 < |x a| < = |c c| < (1)
It is clear in this trivial situation that the condition |c c| < will
hold no matter the choice of > 0; therefore, choose = 1. Thus, for
that choice of , obviously, (1) holds.
Section 9: Presentation of the Theory
Theorem 9.2. For any number a,
lim
xa
x = a. (Rule 2)
Proof : Within the context of the denition of limit of a function,
f(x) = x and L = a. Let > 0. We want to nd a > 0 such that
0 < |x a| < = |f(x) L| <
or,
0 < |x a| < = |x a| < (2)
Towards that end, choose = , then
0 < |x a| < = |x a| < = .
But this is exactly what we wanted to prove, (2).
Section 9: Presentation of the Theory
Theorem 9.3. (Algebra of Limits Theorem) Let f and g be functions
and let a and c be number. Suppose
lim
xa
f(x), and lim
xa
g(x)
exist and are nite. Then,
(1) lim
xa
(f(x) + g(x)) = lim
xa
f(x) + lim
xa
g(x);
(2) lim
xa
(cf(x)) = c lim
xa
f(x);
(3) lim
xa
(f(x)g(x)) = lim
xa
f(x) lim
xa
g(x);
(4) lim
xa
f(x)
g(x)
=
lim
xa
f(x)
lim
xa
g(x)
, provided, lim
xa
g(x) = 0.
Throughout the proofs below, let L = lim
xa
f(x) and M = lim
xa
g(x).
Proof of (1). Let > 0. We must nd a > 0 such that
0 < |x a| < = |(f(x) + g(x)) (L + M)| < .
Section 9: Presentation of the Theory
Towards that end, there exists a
2
> 0 such that
0 < |x a| <
1
= |f(x) L| <

2
, (3)
since lim
xa
f(x) = L. And there exists a
1
> 0 such that
0 < |x a| <
2
= |g(x) M| <

2
, (4)
Finally, dene = min{
1
,
2
}. Note that
0 < |x a| < = 0 < |x a| <
1
and 0 < |x a| <
2
(5)
Now suppose x is a number in the domains of both f and g such that
0 < |x a| < . (6)
Then,
Section 9: Presentation of the Theory
|(f(x) + g(x)) (L + M)|
= |(f(x) L) + (g(x) M)|
|f(x) L| +|g(x) M| Abs. (1)
<

2
+

2
(7)
= .
Thus,
|(f(x) + g(x)) (L + M)| <
Since we are assuming (6), the inequalities on the right-hand side of
(5). Since those inequalities are obtained, then the right-hand sides of
(3) and (4) are true as well. This is were we obtained the inequality
in (7).
We have shown that
0 < |x a| < = |(f(x) + g(x)) (L + M)| < .
This is the denition of lim
xa
(f(x) + g(x)) = L + M.
Section 9: Presentation of the Theory
Proof of (2). Let > 0. We need to nd a number > 0 such that
0 < |x a| < = |(cf(x)) cL| < . (8)
Now, since lim
xa
f(x) = L, and > 0 has been given, from the denition
of limit, there exists a > 0 such that
0 < |x a| < = |f(x) L| <

1 +|c|
. (9)
Now suppose,
0 < |x a| <
then,
|(cf(x)) cL| = |c||f(x) L| Abs. (3)
< |c|

1 +|c|
=
|c|
1 +|c|

.
Thus,
Section 9: Presentation of the Theory
|(cf(x)) cL| <
Note: We utilized the fact that
|c|
1 +|c|
1, for any number c R.
The denominator, 1 +|c|, was chosen instead of |c| to account for the
possibility that c might be zero.
Thus, we have shown that for the > 0 produced in (9), (8) is ob-
tained.
Proof of (3). Let > 0 be given. We want to nd a > 0 such that
0 < |x a| < = |f(x)g(x) LM| < . (10)
Consider the following series of manipulations:
|f(x)g(x) LM|
= |f(x)g(x) f(x)M + f(x)M LM|
|f(x)g(x) f(x)M| +|f(x)M LM| Abs. (1)
= |f(x)||g(x) M| +|M||f(x) L|. Abs. (3) (11)
Section 9: Presentation of the Theory
We can make the second term in (11) small since lim
xa
f(x) = L:
Choose
1
> 0 such that
0 < |x a| <
1
= |f(x) L| <

2(1 +|M|
. (12)
Now for the problem of the rst term of (11). We can make the factor
|g(x) M| as small as we wish since lim
xa
g(x) = M but we have
to make sure that the smallest of this factor is not counter-balanced
by the factor |f(x)|. (That factor might be large we have to make
sure that it is not.) To that end, for 0 < |x a| <
1
, we have, by
(12),
|f(x) L| <

2(1 +|M|)
then,
|f(x)| = |f(x) L + L| |f(x) L| +|L| Abs. (1)

2(1 +|M| +|L|)
Section 9: Presentation of the Theory
For the purpose of convenience, let C =

2(1 +|M| +|L|)
. Thus, there
is a constant, C, such that
0 < |x a| <
1
= |f(x)| < C. (13)
Continuing our quest for the ultimate > 0 for which (10) is valid,
choose
2
> 0 such that
0 < |x a| <
2
= |g(x) M| <

2(1 + C)
. (14)
This is possible since lim
xa
g(x) = M.
Finally, choose the ultimate -value as
= min{
1
,
2
}, (15)
and suppose 0 < |xa| < . Then 0 < |xa| <
1
and 0 < |xa| <
2
form (15) and, as a consequence, the inequalities in (12), (14) and (13)
Section 9: Presentation of the Theory
are valid. Thus, for 0 < |x a| < , we have,
|f(x)g(x) LM|
|f(x)||g(x) M| +|M||f(x) L| (11)
< C

2(1 + C)
+|M|

2(1 +|M|)
(13), (14), (12)
<
C
1 + C

2
+
|M|
1 +|M|

2
.
Thus, we have shown that for any > 0, there exists a > 0 such
that
0 < |x a| < = |f(x)g(x) LM| < .
This completes the proof of this part.
Proof of (4). Let > 0. From the denition, we want to nd a number
> 0 such that
0 < |x a| < =

f(x)
g(x)

L
M

< ,
Section 9: Presentation of the Theory
here, we are assuming M = 0.
Towards this end, consider the following:

f(x)
g(x)

L
M

f(x)M g(x)L
Mg(x)

(f(x)M LM) + (LM g(x)L)

Mg(x)

M(f(x) L) + L(M g(x))

Mg(x)

M(f(x) L)
Mg(x)

L(M g(x))
Mg(x)

Abs. (1)
=
1
|g(x)|
|f(x) L| +
|L|
|Mg(x)|
|g(x) M| Abs. (3)
Do you see where I am going with this calculation? To summarize, we
have shown that

f(x)
g(x)

L
M

1
|g(x)|
|f(x) L| +
|L|
|Mg(x)|
|g(x) M| (16)
Section 9: Presentation of the Theory
Now, for some details. The |g(x)| in the denominator is bothersome.
Since lim
xa
g(x) = M = 0, there is a
1
> 0 such that
0 < |x a| <
1
= |g(x) M| <
M
2
denition
Now, for 0 < |x a| <
1
,
|g(x)| = |M (M g(x))|
|M| |M g(x)| Abs. (2)
|M|
M
2

M
2
.
This establishes the inequality,
0 < |x a| <
1
= |g(x)|
M
2
=
1
|g(x)|

2
M
(18)
Section 9: Presentation of the Theory
Lets put this back into the inequality (16) to get

f(x)
g(x)

L
M

1
|g(x)|
|f(x) L| +
|L|
|Mg(x)|
|g(x) M|

2
|M|
|f(x) L| +
2|L|
M
2
|g(x) M| (20)
Now for our nal calculations! Since lim
xa
f(x) = L, there is a
2
> 0
such that
0 < |x a| <
2
= |f(x) L| <
|M|
4
. (21)
Since lim
xa
g(x) = M, there is a
3
> 0 such that
0 < |x a| <
3
= |g(x) M| <
M
2
4(1 +|L|)
. (22)
Section 9: Presentation of the Theory
Choose = min{
1
,
2
,
3
}. Consequently, if x satises the inequality
0 < |x a| < , then all three of the inequalities (18), (21), and (22)
hold. Thus, if we take x to satisfy,
0 < |x a| < ,
then, from (20), we have

f(x)
g(x)

L
M

2
|M|
|f(x) L| +
2|L|
M
2
|g(x) M|

2
|M|
|M|
4
+
2|L|
M
2
M
2
4(1 +|L|)
from (21) and (22)

2
+

2
=
That does it! We have found the > 0 that works.
Section 9: Presentation of the Theory
Theorem 9.4. (Continuity of Power Functions) Let a R and n
N. Then
lim
xa
x
n
= a
n
. (23)
Proof : By Theorem 9.2, we have lim
xa
x = a. Therefore, by Theo-
rem 9.3 (3),
lim
xa
x
2
= lim
xa
xx = lim
xa
x lim
xa
x = aa = a
2
.
Similarly, by Theorem 9.3 (3),
lim
xa
x
2
= lim
xa
x
2
x = lim
xa
x
2
lim
xa
x = a
2
a = a
3
.
The formal mechanism for nishing the proof in Mathemtical Induc-
tion: Suppose we have shown that
lim
xa
x
n1
= a
n1
,
then, from Theorem 9.3 (3) and Theorem 9.2,
lim
xa
x
n
= lim
xa
x
n1
x = lim
xa
x
n1
lim
xa
x = a
n1
a = a
n
.
Section 9: Presentation of the Theory
By the Principle of Mathematical Induction, we have proved the the-
orem.
Theorem 9.5. (Continuity of Polynomial Functions) Let p be a poly-
nomial and a R. Then
lim
xa
p(x) = p(a). (24)
Proof : Let p(x) be a polynomial of degree n N. This means that
the functional form of p is
p(x) = b
0
+ b
1
x + b
2
x
2
+ b
3
x
3
+ + b
n1
x
n1
+ b
n
x
n
,
for some set of coecients b
0
, b
1
, b
2
, . . . , b
n
.
Notice that p(a)b
0
+ b
1
a + b
2
a
2
+ b
3
a
3
+ + b
n1
a
n1
+ b
n
a
n
.
Section 9: Presentation of the Theory
Now apply Theorem 9.3 (1), Theorem 9.3 (2) as well as Theorem 9.4
to obtain
lim
xa
p(x) = lim
xa
(b
0
+ b
1
x + b
2
x
2
+ b
3
x
3
+ + b
n1
x
n1
+ b
n
x
n
)
= lim
xa
b
0
+ lim
xa
b
1
x + lim
xa
b
2
x
2
+ + lim
xa
b
n
x
n
(1)
= b
0
+ b
1
lim
xa
x + b
2
lim
xa
x
2
+ + b
n
lim
xa
x
n

(2)
Rule 1
= b
0
+ b
1
a + b
2
a
2
+ + b
n
a
n
Thm 9.4
= p(a)
Theorem 9.6. (Continuity of Rational Functions) Let f be a rational
function, and let a Dom(f). Then
lim
xa
f(x) = f(a). (25)
Section 9: Presentation of the Theory
Proof : A rational function is a quotient of two polynomials. Let p and
q be two polynomials such that
f(x) =
p(x)
q(x)
.
The natural domain of f is given by
Dom(f) = { x R | g(x) = 0 }
Let a Dom(f); thus, g(a) = 0. This observation is important as we
are about to cite Theorem 9.3 (4). In that theorem, we require the
limit of the denominator to be nonzero. Read on.
lim
xa
f(x) = lim
xa
p(x)
q(x)
=
lim
xa
p(x)
lim
xa
q(x)
Thm 9.3 (4)
=
p(a)
q(a)
Thm 9.5
Section 9: Presentation of the Theory
Thus,
lim
xa
f(x) =
p(a)
q(a)
= f(a).
This is equation (25) that was asserted in the theorem.
Theorem 9.7. Let f and g be functions that are compatible for com-
position, let a R. Suppose,
(1) lim
xa
g(x) exists, let b = lim
xa
g(x);
(2) b Dom(f), and lim
yb
f(y) = f(b) exists.
Then
lim
xa
f(g(x)), exists
and,
lim
xa
f(g(x)) = f(b),
or,
lim
xa
f(g(x)) = f( lim
xa
g(x)). (26)
Section 9: Presentation of the Theory
Proof : Let > 0. We want to nd a number > 0 such that
0 < |x a| < = |f(g(x)) f(b)| < . (27)
To this end, since > 0 has been given, and it is assumed that
lim
yb
f(y) = f(b), there is a > 0 such that
0 < |y b| < = |f(y) f(b)| < . (28)
Now, we have a number > 0 dened, since we are assuming that
lim
xa
g(x) = b, there is a number > 0 such that
0 < |x a| < = |g(x) b| < . (29)
Now we claim that the number, > 0, produced in the last paragraph
is the -value we seek. Indeed, suppose
0 < |x a| <
Then, from (29),
|g(x) b| < .
Section 9: Presentation of the Theory
But now, this inequality implies, by (28), that
|f(g(x)) f(b)| < .
This proves the theorem!
Theorem 9.8. (Continuity of the Root Function) Let n N. Dene
f(x) =
n

x, for a Dom(f). Then

lim
xa
f(x) = f(a)
or,
lim
xa
n

x =
n

a (30)
Proof : The key to this proof is the observation that f(x) =
n

x is an
increasing function; i.e.,
x
1
, x
2
Dom(f) and x
1
< x
2
= f(x
1
) < f(x
2
)
or,
x
1
, x
2
Dom(f) and x
1
< x
2
=
n

x
1
<
n

x
2
Section 9: Presentation of the Theory
Let a Dom(f). Show lim
xa
n

x =
n

a.
Case I : a = 0. Let > 0, choose =
n
. Then it is obvious that, since
n

0 = 0,
x Dom(f) and |x| < =
n
= |
n

x| <
This is the trivial case.
Case II : a > 0. Let > 0. We can assume is so small that
n

a > 0.
(Why?) Then,
0 <
n

a <
n

a <
n

a + = 0 < (
n

a )
n
< a < (
n

a + )
n
This is because the function x x
n
is increasing on the interval
( 0, ); i.e., 0 < x
1
< x
2
= x
n
1
< x
n
2
.
Dene the > 0 we are looking for as
:= min{ a (
n

a )
n
, (
n

a + )
n
a }
Section 9: Presentation of the Theory
Now suppose 0 < |x a| < . Notice that such an x must necessarily
belong to the domain of f. (Why?) Then
0 < |x a| < = a < x < a +
= (
n

a )
n
< a < x < a + < (
n

a + )
n
(32)
Thus, we have,
0 < |x a| < = (
n

a )
n
< x < (
n

a + )
n
.
Now we use the increasing property of
n

n:
0 < |x a| < = (
n

a )
n
< x < (
n

a + )
n
=
n
_
(
n

a
n
) <
n

x <
n
_
(
n

a +
n
)
=
n

a <
n

x <
n

a +
= |
n

x
n

a| < .
Thus, we have shown that
0 < |x a| < = |
n

x
n

a| < (33)
Section 9: Presentation of the Theory
This proves that lim
xa
n

x =
n

a in this case.
Case III : a < 0. This case is only present with n, the root being
extracted, is an odd integer. The proof is left to the reader it is
similar to Case II. The student need only study the previous case,
make appropriate changes in the steps.
Proof Notes: How can we assume that is such that
n

a > 0? If
does not satisfy this inequality, then choose another -value, say
1
that does satisfy the desired inequality. Now use
1
throughout the
rest of the proof instead of . We would nished the proof with the
declaration that
0 < |x a| < = |
n

x
n

a| <
1
,
but since
1
< , we would have had
0 < |x a| < = |
n

x
n

a| <
1
< ,
which is (33). That being the case, we might as well assume at the
beginning that a > 0 and avoid the introduction of
1
.
Section 9: Presentation of the Theory
If you go to the trouble of studying the proof given in Case II,
it would appear that the only property of
n

x used in the proof was

that it was increasing. (I used that fact that x x
n
is increasing as
well.) That being the case, can this proof be modied to argue that
any increasing function f has the property that lim
xa
f(x) = f(a)?
Having read the previous paragraph, now consider the function
f(x) =
_
x x < 0
1 + x x 0
This function is strictly increasing over R, but lim
x0
f(x) = f(0)
since the two-sided limit does not exist?
Can you resolve the seeming contradiction between the two
paragraphs? e-mail me with your thoughts.
Exercise 9.1. Draw the graph of f(x) =
n

x, and use it to illustrate

the main idea of the proof of Case II of Theorem 9.8.
Exercise 9.2. Let n N be odd, and let a R be negative. Prove
lim
xa
n

x =
n

a.
Section 9: Presentation of the Theory
Corollary 9.9. Suppose lim
xa
g(x) exists, then
lim
xa
n
_
g(x) =
n

lim
xa
g(x), (34)
provided that the number b := lim
xa
g(x) is within the domain of the
n
th
-root function.
Proof : All the heavy lifting has been done. We apply Theorem 9.7
with the function f in that theorem as f(x) =
n

x and the function

g in that theorem, the function g in this corollary! Now the point of
Theorem 9.8 was that it is a verication of condition (2). Of course,
condition (1) is apart of the assumptions of this corollary. Therefore,
we can conclude by Theorem 9.7 that
lim
xa
n
_
g(x) =
n

lim
xa
g(x),
which is (34).
Section 9: Presentation of the Theory
Theorem 9.10. (Continuity of Algebraic Functions) Let f be an al-
gebraic function, and let a Dom(f). Then
lim
xa
f(x) = f(a).
Proof : An algebraic function is constructed by sums, dierences, prod-
ucts, quotients and compositions with functions of the form:
y = c y = x
n
y =
m

x.
These are constant functions, power functions, and root functions,
respectively. For all three, we have shown the property:
lim
xa
c = c lim
xa
x
n
= a
n
lim
xa
m

x =
m

a.
The latter case is true provided a belongs to the domain of the m
th
-
root functions. (See theorems: 9.1, 9.4, and 9.8.)
These observations, combined with the Algebra of Limits Theorem,
which concerns sums, dierences, products, and quotients of functions
Section 9: Presentation of the Theory
whose limits exist, and with Theorem 9.7, which concerns composi-
tions of functions whose limits exist, allows us to make the assertion
of the theorem.
Theorem 9.11. Let g, f, and h be functions and a, L R. Suppose
there is some
0
> 0 such that
g(x) f(x) h(x) |x a| <
0
, (35)
and,
lim
xa
g(x) = lim
xa
h(x) = L.
Then
lim
xa
f(x) = L.
Proof : Let > 0.
Since lim
xa
g(x) = L, there is a
1
> 0 such that
0 < |x a| <
1
= |g(x) L| < .
Section 9: Presentation of the Theory
This inequality implies,
L < g(x) whenever, 0 < |x a| <
1
(36)
(Do you know why?)
Now, since lim
xa
h(x) = L, there is a
2
> 0 such that
0 < |x a| <
2
= |h(x) L| < .
This implies
h(x) < L + whenever, 0 < |x a| <
2
(37)
(Do you know why?)
Finally, choose = min{
0
,
1
,
2
}, then
0 < |x a| <
implies,
L <
. .
(36)
g(x) f(x) h(x)
. .
(35)
< L +
. .
(37)
Section 9: Presentation of the Theory
But this implies,
|f(x) L| < , whenever 0 < |x a| < .
which is what we wanted to prove.
Theorem 9.12. The following limits are obtained.
lim
x0
sin(x) = 0 lim
x0
cos(x) = 1. (38)
Proof : Under construction.
Theorem 9.13. The following limits are obtained.
lim
x0
sin(x)
x
= 1 lim
x0
1 cos(x)
x
= 0. (39)
Proof : Under construction.
Appendix
Properties of Absolute Value. Let a, b, c R, then each of the
following inequalities are obtained:
1. |a + b| |a| +|b|.
2. |a b| |a| |b|.
3. |ab| = |a||b|.
Solutions to Exercises
9.1. Draw a graph that represents f. Designate a point on the hori-
zontal axis as a, then mark o the corresponding point on the vertical
axis label this point
n

a.
Now mark o a little distance, , equidistant above and below the
point labeled
n

a. These are the points

n

a and
n

a + . Label
them so.
Now starting at each of the points
n

a and
n

a + , move hori-
zontally until you hit the graph of f, now move vertically downward
until you intersect the x-axis. The two points obtained in this way
are (
n

a )
n
and (
n

a +)
n
, respectively. These two points have the
number a between them.
The quantity is the shortest distance between a and each of the
two points constructed in the previous paragraph. Now mark o a
-distance on either side of a. Do you see that this interval lies com-
pletely inside the larger interval? (Because is the shortest distance
Solutions to Exercises (continued)
to the endpoints.) Now it is clear, at least geometrically, the validity
of the inequality (32). Can you show the inequality (32) algebraically?
Assuming you understand my verbal (bit-tell?) instructions, do you
see why
0 < |x a| < = |
n

x
n

a| < ?
Exercise Notes: Review my question in the Proof Notes above. Does
this graphical construction give you additional insight into answering
the question?
Exercise 9.1.
9.2. I said it is left to the reader! DPS
Study the proof of Case II of Theorem 9.8 and make appropriate
changes to correspond to a < 0.
Exercise 9.2.